The effect of a dominant shrub on annuals plants in a desert system changes along a gradient of precipitation.
Alex Filazzola
May 2017
Ephedra regional gradient (ERG) analyses
Abstract
An under-examined component of the shrub-annual relationship is how regional drivers, such as climate, may alter the sign or magnitude of positive interactions. Interspecific interactions between plants have been shown to be strongly linked to climate, particularly temperature and precipitation. The stress-gradient hypothesis (SGH) predicts that higher abiotic stress (i.e. for deserts, increasing temperature and reduced precipitation) will increase the frequency of positive interactions among shrubs and their annual understory. Regional climate gradients also have indirect effects on plant composition, such as determining consumer abundance and soil nutrient composition. Nutrient availability is particularly affected by precipitation because of altered decomposition rates of organic matter and mineralization. Therefore, the strength of facilitation and operating mechanism of a shrub on the annual plant community may change along a regional gradient.
Hypothesis
We tested the hypothesis that positive interactions among shrubs and annual plants will increase with abiotic stress and reduce nutrient availability along a regional gradient of aridity.
Methods
load GAMs
Year<-c("2016","2016","2016","2016","2016","2016","2016","2017","2017","2017","2017","2017","2017","2017")
Site<-c("Barstow","Cuyama","HeartofMojave","PanocheHills","SheepholeValley","Tecopa","TejonRanch","Barstow","Cuyama","HeartofMojave","PanocheHills","SheepholeValley","Tecopa","TejonRanch")
gamsq <-c(7992.36,6789.76,7903.21,5715.36,7779.24,7482.25,7447.69,6872.41,5882.89,7157.16,5595.04,7293.16,7106.49,6955.56)
gams.richard <-c(89.4,82.4,88.9,75.6,88.2,86.5,86.3,82.9,76.7,84.6,74.8,85.4,84.3,83.4)
gams <- c(89.6,71,84.7,62.7,88.3,80.6,80.7,63.6,51.4,73.4,49.5,69.4,68.3,65.8)
gams.data <- data.frame(Year,Site,gamsq,gams)Weather for growing season
Climate patterns within study
season1.sjd <- season1 %>% filter(Gradient<4) %>% group_by(year, month,days) %>% summarise_if(is.numeric, funs(mean(., na.rm=T)))
season1.mnp <- season1 %>% filter(Gradient>3)%>% group_by(year, month,days) %>% summarise_if(is.numeric, funs(mean(., na.rm=T)))
season2.sjd <- season2 %>% filter(Gradient<4) %>% group_by(year, month,days) %>% summarise_if(is.numeric, funs(mean(., na.rm=T)))
season2.mnp <- season2 %>% filter(Gradient>3)%>% group_by(year, month,days) %>% summarise_if(is.numeric, funs(mean(., na.rm=T)))
## Rain vs Temperature in 2016
par(mfrow=c(2,1))
par(mar=c(3.5,4.5,1,4.5))
plot1 <- barplot(height=season1.sjd$Precip, ylim=c(0,14), ylab="Average precipitation San Joaquin (cm)")
points(plot1[,1], season1.sjd$min.temp, type="l", col="#FF000050", lwd=2)
axis(4, at=seq(0,14,2), lab=seq(0,14,2), ylab="")
mtext("Average temperature at all sites (°C)", 4, line=3)
par(mar=c(4.5,4.5,0,4.5))
plot1 <- barplot(height=season1.mnp$Precip, ylim=c(0,14), ylab="Average precipitation Mojave (cm)")
axis(1, plot1[c(1,30,60,90,120,150,180)], c("Nov","Dec","Jan","Feb","Mar","Apr","May"))
points(plot1[,1], season1.mnp$min.temp, type="l", col="#FF000050", lwd=2)
axis(4, at=seq(0,14,2), lab=seq(0,14,2), ylab="")
mtext("Average temperature at all sites (°C)", 4, line=3)
## Rain vs Temperature in 2017
par(mfrow=c(2,1))
par(mar=c(3.5,4.5,1,4.5))
plot2 <- barplot(height=season2.sjd$Precip, ylim=c(0,14), ylab="Average precipitation San Joaquin (cm)")
points(plot2[,1], season2.sjd$min.temp, type="l", col="#FF000050", lwd=2)
axis(4, at=seq(0,14,2), lab=seq(0,14,2), ylab="")
mtext("Average temperature San Joaquin (°C)", 4, line=3)
par(mar=c(4.5,4.5,0,4.5))
plot2 <- barplot(height=season2.mnp$Precip, ylim=c(0,14), ylab="Average precipitation Mojave (cm)")
points(plot2[,1], season2.mnp$min.temp, type="l", col="#FF000050", lwd=2)
axis(1, plot1[c(1,30,60,90,120,150,180)], c("Nov","Dec","Jan","Feb","Mar","Apr","May"))
axis(4, at=seq(0,14,2), lab=seq(0,14,2), ylab="")
mtext("Average temperature Mojave (°C)", 4, line=3)
### 2016 The rain was inconsistent and mostly absent in the Mojave. This resulted in low germination and producitivty at the southern sites
### 2017 The rain was more plentiful, but in the northern sites, there appears to be a frost period after the majority of the rainfall. Need to check number of frost days
season1.frost <- season1 %>% group_by(Site) %>% summarize(frost.days=sum(min.temp<0, na.rm=T)/length(min.temp)*100)
data.frame(season1.frost)## Site frost.days
## 1 Barstow 24.725275
## 2 Cuyama 29.670330
## 3 HeartofMojave 25.824176
## 4 PanocheHills 10.439560
## 5 SheepholeValley 6.043956
## 6 Tecopa 23.626374
## 7 TejonRanch 50.549451season2.frost <- season2 %>% group_by(Site) %>% summarize(frost.days=sum(min.temp<0, na.rm=T)/length(min.temp)*100)
data.frame(season2.frost)## Site frost.days
## 1 Barstow 17.679558
## 2 Cuyama 19.889503
## 3 HeartofMojave 16.574586
## 4 PanocheHills 14.917127
## 5 SheepholeValley 1.785714
## 6 Tecopa 25.966851
## 7 TejonRanch 35.911602## Both years had comparable number of frost days
## Compare number of consecutive frost days (i.e. frost periods)
season1[,"frost"] <- ifelse(season1$min.temp<0, -99,season1$min.temp) ## identified days below freezing
season2[,"frost"] <- ifelse(season2$min.temp<0, -99,season2$min.temp) ## identified days below freezing
count.consec <- function(x) {max(rle(as.character(x))$lengths)}
season1.frost <- season1 %>% group_by(Site) %>% summarize(count.consec(frost))
data.frame(season1.frost)## Site count.consec.frost.
## 1 Barstow 10
## 2 Cuyama 10
## 3 HeartofMojave 12
## 4 PanocheHills 5
## 5 SheepholeValley 7
## 6 Tecopa 11
## 7 TejonRanch 14season2.frost <- season2 %>% group_by(Site) %>% summarize(count.consec(frost))
data.frame(season2.frost)## Site count.consec.frost.
## 1 Barstow 5
## 2 Cuyama 6
## 3 HeartofMojave 7
## 4 PanocheHills 6
## 5 SheepholeValley 3
## 6 Tecopa 7
## 7 TejonRanch 13## compare only after plants have germinated
season1.frost <- season1 %>% group_by(Site) %>% filter(year>2015) %>% summarize(frost.days=sum(min.temp<0, na.rm=T)/length(min.temp)*100, avg.min.temp=mean(min.temp, na.rm=T))
data.frame(season1.frost)## Site frost.days avg.min.temp
## 1 Barstow 12.396694 5.750413
## 2 Cuyama 11.570248 3.352893
## 3 HeartofMojave 16.528926 4.542149
## 4 PanocheHills 2.479339 6.777686
## 5 SheepholeValley 2.479339 9.394167
## 6 Tecopa 11.570248 6.342149
## 7 TejonRanch 33.884298 1.438017season2.frost <- season2 %>% group_by(Site) %>% filter(year>2016) %>% summarize(frost.days=sum(min.temp<0, na.rm=T)/length(min.temp)*100,avg.min.temp=mean(min.temp, na.rm=T))
data.frame(season2.frost)## Site frost.days avg.min.temp
## 1 Barstow 11.666667 6.052500
## 2 Cuyama 16.666667 3.624167
## 3 HeartofMojave 8.333333 5.637838
## 4 PanocheHills 8.333333 6.065833
## 5 SheepholeValley 0.000000 9.386916
## 6 Tecopa 20.833333 5.938333
## 7 TejonRanch 28.333333 2.535833season1.frost <- season1 %>% group_by(Site) %>% filter(year>2015) %>% summarize(count.consec(frost))
data.frame(season1.frost)## Site count.consec.frost.
## 1 Barstow 5
## 2 Cuyama 6
## 3 HeartofMojave 4
## 4 PanocheHills 2
## 5 SheepholeValley 2
## 6 Tecopa 4
## 7 TejonRanch 10season2.frost <- season2 %>% group_by(Site) %>% filter(year>2016)%>% summarize(count.consec(frost))
data.frame(season2.frost)## Site count.consec.frost.
## 1 Barstow 5
## 2 Cuyama 6
## 3 HeartofMojave 3
## 4 PanocheHills 3
## 5 SheepholeValley 2
## 6 Tecopa 5
## 7 TejonRanch 10Microenvironmental differences
We compared temperature and relative humidity between shrub and open microsites among all sites along the regional gradient 
Shapiro-Wilk normality test
data: fit1$residuals
W = 0.87899, p-value < 2.2e-16 Df Sum Sq Mean Sq F value Pr(>F)
Microsite 1 33.6 33.59 76.013 <2e-16 ***
gradient 6 234.0 39.00 88.248 <2e-16 ***
Microsite:gradient 6 8.8 1.46 3.306 0.003 **
Residuals 4290 1896.0 0.44
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
30 observations deleted due to missingnessAnalysis of Deviance Table
Model: binomial, link: logit
Response: RH/100
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 4333 974.99
Microsite 1 2.11 4332 972.88 0.1463
gradient 6 528.28 4326 444.60 <2e-16 ***
Microsite:gradient 6 7.44 4320 437.17 0.2824
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Analysis of Deviance Table
Model: Gamma, link: inverse
Response: swc
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev F Pr(>F)
NULL 404 231.405
Site 6 189.683 398 41.722 381.2179 <2e-16 ***
Microsite 1 0.287 397 41.435 3.4579 0.0637 .
Site:Microsite 6 0.696 391 40.739 1.3994 0.2136
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Analysis of Deviance Table
Model: Gamma, link: inverse
Response: swc
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev F Pr(>F)
NULL 419 183.193
Site 6 144.930 413 38.263 279.4537 <2e-16 ***
Microsite 1 0.080 412 38.183 0.9284 0.3358
Site:Microsite 6 0.535 406 37.647 1.0322 0.4037
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
site level means
## site coordintes
site.gps <- read.csv("Data/ERGsites.csv")
# nutrient data for each site
nutrients <- read.csv("Data/ERG.soilnutrients.csv")
## long term climate data extracted from worldclim
clim.dat <- read.csv("Data/ERG.worldclim.csv")
## check correlation among long-term climate variables
cor(clim.dat[,3:8]) ## Temp.wettest.QR least correlated ## Annual.Temp Temp.Seasonality Temp.wettest.QR
## Annual.Temp 1.0000000 0.9051791 0.7260617
## Temp.Seasonality 0.9051791 1.0000000 0.4650548
## Temp.wettest.QR 0.7260617 0.4650548 1.0000000
## Annual.precip -0.8911069 -0.9650533 -0.5313460
## Precip.seasonality -0.9715419 -0.9774646 -0.6165170
## Precipt.wettest.QR -0.8839124 -0.9570957 -0.5523347
## Annual.precip Precip.seasonality Precipt.wettest.QR
## Annual.Temp -0.8911069 -0.9715419 -0.8839124
## Temp.Seasonality -0.9650533 -0.9774646 -0.9570957
## Temp.wettest.QR -0.5313460 -0.6165170 -0.5523347
## Annual.precip 1.0000000 0.9613729 0.9980959
## Precip.seasonality 0.9613729 1.0000000 0.9543657
## Precipt.wettest.QR 0.9980959 0.9543657 1.0000000## Obtain mean shrub traits for each site
shrubs <- read.csv("Data/ERG.shrub.csv")
shrubs <- subset(shrubs, Microsite=="shrub")
shrubs.mean <- shrubs %>% group_by(Gradient,Site) %>% summarise_all(funs(mean))
shrubs.mean <- data.frame(shrubs.mean)
shrubs.vars <- shrubs.mean[,c("Site","Gradient","volume","canopy","Dx","DxEph","Compaction")]
## test correlation among shrub traits
cor(shrubs.vars[,3:7]) ## compaction x volume & DxEph x Dx high correlation## volume canopy Dx DxEph Compaction
## volume 1.0000000 0.2243902 0.4003665 0.2631563 -0.7023062
## canopy 0.2243902 1.0000000 -0.6847957 -0.3538094 -0.1437590
## Dx 0.4003665 -0.6847957 1.0000000 0.7803527 -0.1355700
## DxEph 0.2631563 -0.3538094 0.7803527 1.0000000 0.3221944
## Compaction -0.7023062 -0.1437590 -0.1355700 0.3221944 1.0000000vifstep(shrubs.vars[,3:7], th=10) ## Dx showing collinearity problems## 1 variables from the 5 input variables have collinearity problem:
##
## Dx
##
## After excluding the collinear variables, the linear correlation coefficients ranges between:
## min correlation ( Compaction ~ canopy ): -0.143759
## max correlation ( Compaction ~ volume ): -0.7023062
##
## ---------- VIFs of the remained variables --------
## Variables VIF
## 1 volume 7.026144
## 2 canopy 1.925499
## 3 DxEph 4.316058
## 4 Compaction 6.400914shrub.site <- shrubs.vars[,-c(1,2,5)]
rownames(shrub.site) <- shrubs.vars[,1]
## PCA of shrub characteristics
pca1 <- prcomp(log(shrub.site), scale=T)
plot(pca1)
biplot(pca1, scale = T)
summary(pca1) ## 77% variation explained## Importance of components:
## PC1 PC2 PC3 PC4
## Standard deviation 1.3017 1.1778 0.9082 0.3059
## Proportion of Variance 0.4236 0.3468 0.2062 0.0234
## Cumulative Proportion 0.4236 0.7704 0.9766 1.0000## PCA of site characteristics for season1
## Obtain mean weather variables for each site
season1.mean <- season1 %>% group_by(Gradient,Site) %>% summarise(temp.var=var(avg.temp,na.rm=T),Precip=sum(Precip),wind=mean(wind.speed, na.rm=T))
season1.mean <- data.frame(season1.mean)
## extract key variables
## dropped RH, and chose min.temp because least correlation with others & cold stress
##combine nutrients and long-term averages
site.vars <- data.frame(season1.mean[,3:5],site.gps["elevation"]) ## drop Phosphorus and other climate variables because of correlations,
row.names(site.vars) <- shrubs.vars[,1]
cor(site.vars)## temp.var Precip wind elevation
## temp.var 1.00000000 -0.8679054 0.05816084 -0.2699588
## Precip -0.86790537 1.0000000 -0.37279135 0.1222504
## wind 0.05816084 -0.3727914 1.00000000 0.1054497
## elevation -0.26995879 0.1222504 0.10544974 1.0000000site.vars2016 <- site.vars
## check for collinearity
vifstep(site.vars, th=10) ## remove potassium and temperature minimum## No variable from the 4 input variables has collinearity problem.
##
## The linear correlation coefficients ranges between:
## min correlation ( wind ~ temp.var ): 0.05816084
## max correlation ( Precip ~ temp.var ): -0.8679054
##
## ---------- VIFs of the remained variables --------
## Variables VIF
## 1 temp.var 6.640710
## 2 Precip 7.316232
## 3 wind 1.739558
## 4 elevation 1.142681pca1 <- prcomp(log(abs(site.vars)), scale=T)
plot(pca1)
biplot(pca1)
## check contribution of loadings
pca1$rotation## PC1 PC2 PC3 PC4
## temp.var 0.6001002 -0.1669018 0.50554765 -0.5970303
## Precip -0.6703313 -0.1105034 -0.09936844 -0.7270288
## wind 0.2743589 0.7981791 -0.43410209 -0.3149488
## elevation -0.3395039 0.5681927 0.73898773 0.1256633aload <- abs(pca1$rotation)
sweep(aload, 2, colSums(aload), "/")## PC1 PC2 PC3 PC4
## temp.var 0.3184748 0.1015355 0.28433406 0.33832382
## Precip 0.3557466 0.0672253 0.05588758 0.41199111
## wind 0.1456030 0.4855763 0.24415109 0.17847449
## elevation 0.1801756 0.3456629 0.41562726 0.07121058summary(pca1) ## 89% variation explained## Importance of components:
## PC1 PC2 PC3 PC4
## Standard deviation 1.456 1.0458 0.8577 0.22479
## Proportion of Variance 0.530 0.2734 0.1839 0.01263
## Cumulative Proportion 0.530 0.8034 0.9874 1.00000gradient1.season1 <- pca1$x[,1]
gradient2.season1 <- pca1$x[,2]
## season2
season2.mean <- season2 %>% group_by(Gradient,Site) %>% summarise(temp.var=var(avg.temp,na.rm=T),Precip=sum(Precip, na.rm=T),wind=mean(wind.speed, na.rm=T))
season2.mean <- data.frame(season2.mean)
## extract key variables
## dropped RH, and chose min.temp because least correlation with others & cold stress
##combine nutrients and long-term averages
site.vars <- data.frame(season2.mean[,3:5],site.gps["elevation"]) ## drop Phosphorus and other climate variables because of correlations,
row.names(site.vars) <- shrubs.vars[,1]
cor(site.vars)## temp.var Precip wind elevation
## temp.var 1.0000000 -0.7427028 -0.2262409 -0.3715471
## Precip -0.7427028 1.0000000 -0.1551517 0.1813600
## wind -0.2262409 -0.1551517 1.0000000 0.1052805
## elevation -0.3715471 0.1813600 0.1052805 1.0000000site.vars2017 <- site.vars
## check for collinearity
vifstep(site.vars, th=10) ## remove potassium and temperature minimum## No variable from the 4 input variables has collinearity problem.
##
## The linear correlation coefficients ranges between:
## min correlation ( elevation ~ wind ): 0.1052805
## max correlation ( Precip ~ temp.var ): -0.7427028
##
## ---------- VIFs of the remained variables --------
## Variables VIF
## 1 temp.var 3.438786
## 2 Precip 3.035337
## 3 wind 1.402010
## 4 elevation 1.192010pca2 <- prcomp(log(abs(site.vars)), scale=T)
plot(pca2)
biplot(pca2)
summary(pca2) ## 85% variation explained## Importance of components:
## PC1 PC2 PC3 PC4
## Standard deviation 1.434 1.0121 0.8809 0.37883
## Proportion of Variance 0.514 0.2561 0.1940 0.03588
## Cumulative Proportion 0.514 0.7701 0.9641 1.00000## check contribution of loadings
pca2$rotation## PC1 PC2 PC3 PC4
## temp.var 0.6557532 -0.09123357 0.20565681 0.7206729
## Precip -0.6176537 -0.16891771 -0.40105332 0.6550778
## wind -0.1019884 0.97279862 0.06826009 0.1964733
## elevation -0.4220070 -0.12963831 0.89005734 0.1135866aload <- abs(pca2$rotation)
sweep(aload, 2, colSums(aload), "/")## PC1 PC2 PC3 PC4
## temp.var 0.36483385 0.06695609 0.1314078 0.42749339
## Precip 0.34363685 0.12396828 0.2562596 0.38858328
## wind 0.05674213 0.71393442 0.0436159 0.11654531
## elevation 0.23478717 0.09514122 0.5687167 0.06737801## define gradients
gradient1.season2 <- pca2$x[,1]
gradient2.season2 <- pca2$x[,2]Compare gradient to phytometer
Differences from Nutrient content
## join aridity with nutrients
nutrients.climate <- merge(nutrients,subset(site.climate, Year==2016), by=c("Site"))
## Nitrogen difference between shrub and sites
m.nit <- aov(log(N) ~ Site * microsite, data=nutrients)
summary(m.nit)## Df Sum Sq Mean Sq F value Pr(>F)
## Site 6 63.04 10.51 13.241 2.86e-09 ***
## microsite 1 40.16 40.16 50.614 2.24e-09 ***
## Site:microsite 6 4.96 0.83 1.042 0.408
## Residuals 56 44.43 0.79
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1TukeyHSD(m.nit, "microsite")## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(N) ~ Site * microsite, data = nutrients)
##
## $microsite
## diff lwr upr p adj
## shrub-open 1.514873 1.088317 1.941429 0## Potassium difference between shrub and sites
m.pot <- aov(log(K) ~ Site * microsite, data=nutrients)
summary(m.pot)## Df Sum Sq Mean Sq F value Pr(>F)
## Site 6 15.897 2.650 22.013 4.09e-13 ***
## microsite 1 4.445 4.445 36.934 1.14e-07 ***
## Site:microsite 6 1.605 0.268 2.223 0.0541 .
## Residuals 56 6.740 0.120
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1TukeyHSD(m.pot, "microsite")## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(K) ~ Site * microsite, data = nutrients)
##
## $microsite
## diff lwr upr p adj
## shrub-open 0.5040081 0.3378755 0.6701406 1e-07## Phosphorus difference between shrub and sites
m.pho <- aov(log(P) ~ Site * microsite, data=nutrients)
summary(m.pho)## Df Sum Sq Mean Sq F value Pr(>F)
## Site 6 23.163 3.861 14.33 8.10e-10 ***
## microsite 1 10.546 10.546 39.15 5.79e-08 ***
## Site:microsite 6 3.394 0.566 2.10 0.0676 .
## Residuals 56 15.085 0.269
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1TukeyHSD(m.pho, "microsite")## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(P) ~ Site * microsite, data = nutrients)
##
## $microsite
## diff lwr upr p adj
## shrub-open 0.7762734 0.5277317 1.024815 1e-07nutrient.mean <- nutrients.climate %>% group_by(Site, arid.gradient) %>% summarize(nitrogen=mean(N), phosphorus=mean(P), potassium=mean(K))
nutrient.mean <- data.frame(nutrient.mean)
ggplot(nutrient.mean) + geom_jitter(aes(x=arid.gradient, y=nitrogen), color="#E69F00", size=3) + stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=nitrogen), color="#E69F00") + geom_jitter(aes(x=arid.gradient, y=potassium/20), color="#56B4E9", size=3) + stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=potassium/20), color="#56B4E9") + geom_jitter(aes(x=arid.gradient, y=phosphorus), color="#009E73", size=3) +ylab("soil nutrient content")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) + xlab("aridity gradient")+
annotate("segment", x = -4.5, xend = -4.35, y = 25, yend = 25, colour = "#E69F00", size=2) + ## add custom legend
annotate("text", x = -4.3, y = 25, label = "Nitrogen", size=5, hjust=0)+
annotate("segment", x = -4.5, xend = -4.35, y = 23, yend = 23, colour = "#56B4E9", size=2) + ## add custom legend
annotate("text", x = -4.3, y = 23, label = "Potassium", size=5, hjust=0)+
annotate("segment", x = -4.5, xend = -4.35, y = 21, yend = 21, colour = "#009E73", size=2) + ## add custom legend
annotate("text", x = -4.3, y = 21, label = "Phosphorus", size=5, hjust=0)
m1 <- lm(nitrogen~poly(arid.gradient,2), data=nutrient.mean)
summary(m1)##
## Call:
## lm(formula = nitrogen ~ poly(arid.gradient, 2), data = nutrient.mean)
##
## Residuals:
## 1 2 3 4 5 6 7
## 1.1092 1.0570 -0.4565 -0.8511 -0.8789 -0.3757 0.3960
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 5.4326 0.3942 13.782 0.000161 ***
## poly(arid.gradient, 2)1 -12.6364 1.0429 -12.116 0.000266 ***
## poly(arid.gradient, 2)2 7.0832 1.0429 6.792 0.002454 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.043 on 4 degrees of freedom
## Multiple R-squared: 0.9797, Adjusted R-squared: 0.9695
## F-statistic: 96.47 on 2 and 4 DF, p-value: 0.0004126m2 <- lm(phosphorus~arid.gradient, data=nutrient.mean)
summary(m2)##
## Call:
## lm(formula = phosphorus ~ arid.gradient, data = nutrient.mean)
##
## Residuals:
## 1 2 3 4 5 6 7
## 4.3035 1.1485 -2.0485 1.1930 -0.2791 -5.5020 1.1845
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 14.191 4.527 3.135 0.0258 *
## arid.gradient 1.780 1.298 1.371 0.2287
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3.383 on 5 degrees of freedom
## Multiple R-squared: 0.2732, Adjusted R-squared: 0.1278
## F-statistic: 1.879 on 1 and 5 DF, p-value: 0.2287m3 <- lm(potassium~poly(arid.gradient,2), data=nutrient.mean)
summary(m3)##
## Call:
## lm(formula = potassium ~ poly(arid.gradient, 2), data = nutrient.mean)
##
## Residuals:
## 1 2 3 4 5 6 7
## -3.551 -65.927 32.611 57.492 -25.220 52.524 -47.931
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 186.56 22.69 8.221 0.00119 **
## poly(arid.gradient, 2)1 189.58 60.04 3.157 0.03427 *
## poly(arid.gradient, 2)2 190.31 60.04 3.170 0.03387 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 60.04 on 4 degrees of freedom
## Multiple R-squared: 0.8334, Adjusted R-squared: 0.7502
## F-statistic: 10.01 on 2 and 4 DF, p-value: 0.02774annual plant community repsonse
spp.data <- read.csv("Data/ERG.communitydata.csv")
spp.data[is.na(spp.data)] <- 0
mean.spp <- spp.data %>% group_by(Year,Site, Microsite) %>% summarize(abd=mean(Abundance), richness=mean(Richness), biomass=mean(Biomass))
mean.spp <- data.frame(mean.spp)
## collect aridity gradient data
site.climate <- rbind(s1.mean,s2.mean)
site.climate[,"Year"] <- c(rep("2016",7),rep("2017",7))
## combine climate data with community data
mean.spp <- merge(mean.spp, site.climate, by.y=c("Year","Site"))
## community models with GAMs
community <- spp.data
comm.clim <- merge(community,gams.data, by=c("Year","Site"))
m1 <- glmer.nb(Abundance~ Microsite * poly(gams,2) + (1|Year), data=comm.clim)
car::Anova(m1, type=c("II"))## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: Abundance
## Chisq Df Pr(>Chisq)
## Microsite 19.0016 1 1.306e-05 ***
## poly(gams, 2) 1328.4369 2 < 2.2e-16 ***
## Microsite:poly(gams, 2) 3.1951 2 0.2024
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m2 <- glmer.nb(Richness~ Microsite * poly(gams,2) + (1|Year), data=comm.clim)## Warning in theta.ml(Y, mu, weights = object@resp$weights, limit = limit, :
## iteration limit reachedm2.poi <- glmer(Richness~ Microsite * poly(gams,2) + (1|Year), data=comm.clim, family="poisson")
car::Anova(m2.poi, type=c("III"))## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: Richness
## Chisq Df Pr(>Chisq)
## (Intercept) 135.1771 1 < 2.2e-16 ***
## Microsite 7.8474 1 0.005089 **
## poly(gams, 2) 69.6946 2 7.345e-16 ***
## Microsite:poly(gams, 2) 32.9738 2 6.916e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m3 <- lmer(ihs(Biomass)~ Microsite * poly(gams,2) + (1|Year), data=comm.clim)
anova(m3)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 64.53 64.527 1 833.00 210.49 < 2.2e-16 ***
## poly(gams, 2) 511.61 255.805 2 833.44 834.46 < 2.2e-16 ***
## Microsite:poly(gams, 2) 32.08 16.041 2 833.00 52.33 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# ## community responses season1
# mean.spp2016 <- subset(mean.spp, Year==2016)
#
# ## richness
# plot(mean.spp2016[mean.spp2016$Microsite=="open","arid.gradient"], mean.spp2016[mean.spp2016$Microsite=="open","richness"], ylab="average site richness", xlab="aridity gradient")
# points(mean.spp2016[mean.spp2016$Microsite=="shrub","arid.gradient"],mean.spp2016[mean.spp2016$Microsite=="shrub","richness"], pch=19)
#
# m1 <- lm(richness~arid.gradient*Microsite, data=mean.spp2016)
# summary(m1)
#
# ## abundance
# plot(mean.spp2016[mean.spp2016$Microsite=="open","arid.gradient"], mean.spp2016[mean.spp2016$Microsite=="open","abd"], ylab="average site abundance", xlab="aridity gradient")
# points(mean.spp2016[mean.spp2016$Microsite=="shrub","arid.gradient"],mean.spp2016[mean.spp2016$Microsite=="shrub","abd"], pch=19)
#
# m2 <- lm(abd~arid.gradient*Microsite, data=mean.spp2016)
# summary(m2)
#
# ## biomass
# plot(mean.spp2016[mean.spp2016$Microsite=="open","arid.gradient"], mean.spp2016[mean.spp2016$Microsite=="open","biomass"], ylab="average site biomass", xlab="aridity gradient")
# points(mean.spp2016[mean.spp2016$Microsite=="shrub","arid.gradient"],mean.spp2016[mean.spp2016$Microsite=="shrub","biomass"], pch=19)
#
# m3 <- lm(biomass~arid.gradient*Microsite, data=mean.spp2016)
# summary(m3)
#
# ##season2
#
# ## community responses season2
#
# mean.spp2017 <- subset(mean.spp, Year==2017)
#
#
# ## richness
# plot(mean.spp2017[mean.spp2017$Microsite=="open","arid.gradient"], mean.spp2017[mean.spp2017$Microsite=="open","richness"])
# points(mean.spp2017[mean.spp2017$Microsite=="shrub","arid.gradient"],mean.spp2017[mean.spp2017$Microsite=="shrub","richness"], pch=19)
#
# m1 <- lm(richness~arid.gradient*Microsite, data=mean.spp2017)
# summary(m1)
#
# ## abundance
# plot(mean.spp2017[mean.spp2017$Microsite=="open","arid.gradient"], mean.spp2017[mean.spp2017$Microsite=="open","abd"])
# points(mean.spp2017[mean.spp2017$Microsite=="shrub","arid.gradient"],mean.spp2017[mean.spp2017$Microsite=="shrub","abd"], pch=19)
#
# m2 <- lm(abd~arid.gradient*Microsite, data=mean.spp2017)
# summary(m2)
#
# ## biomass
# plot(mean.spp2017[mean.spp2017$Microsite=="open","arid.gradient"], mean.spp2017[mean.spp2017$Microsite=="open","biomass"])
# points(mean.spp2017[mean.spp2017$Microsite=="shrub","arid.gradient"],mean.spp2017[mean.spp2017$Microsite=="shrub","biomass"], pch=19)
#
# m3 <- lm(biomass~arid.gradient*Microsite, data=mean.spp2017)
# summary(m3)
# ## both seasons
#
# ## richness
# plot(mean.spp[mean.spp$Microsite=="open","arid.gradient"], mean.spp[mean.spp$Microsite=="open","richness"], ylab="average site richness", xlab="aridity gradient")
# points(mean.spp[mean.spp$Microsite=="shrub","arid.gradient"],mean.spp[mean.spp$Microsite=="shrub","richness"], pch=19)
#
# m1 <- lm(richness~arid.gradient*Microsite, data=mean.spp)
# summary(m1)
#
# m1 <- lm(richness~poly(arid.gradient,2), data=mean.spp)
# summary(m1)
#
# ggplot(mean.spp) + geom_jitter(aes(x=arid.gradient, y=richness, fill=Microsite, color=Microsite), size=3) +
# scale_color_manual(values=c("#E69F00","#56B4E9")) + stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=richness), color="black")+ ylab("Species richness") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16))
#
# ## abundance
# plot(mean.spp[mean.spp$Microsite=="open","arid.gradient"], mean.spp[mean.spp$Microsite=="open","abd"], ylab="average site abundance", xlab="aridity gradient")
# points(mean.spp[mean.spp$Microsite=="shrub","arid.gradient"],mean.spp[mean.spp$Microsite=="shrub","abd"], pch=19)
#
# m2 <- lm(abd~arid.gradient*Microsite, data=mean.spp)
# summary(m2)
#
# m2 <- lm(ihs(abd)~arid.gradient, data=mean.spp)
# summary(m2)
#
# ggplot(mean.spp) + geom_jitter(aes(x=arid.gradient, y=ihs(abd), fill=Microsite, color=Microsite), size=3)+
# scale_color_manual(values=c("#E69F00","#56B4E9")) + stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=ihs(abd)), color="black")+ ylab("Annual abundance") + xlab("aridity gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16))
#
#
#
# ## biomass
# plot(mean.spp[mean.spp$Microsite=="open","arid.gradient"], ihs(mean.spp[mean.spp$Microsite=="open","biomass"]), ylab="average site biomass", xlab="aridity gradient")
# points(mean.spp[mean.spp$Microsite=="shrub","arid.gradient"],ihs(mean.spp[mean.spp$Microsite=="shrub","biomass"]), pch=19)
#
# m3 <- lm(ihs(biomass)~arid.gradient*Microsite, data=mean.spp)
# summary(m3)
#
# ggplot(mean.spp) + geom_jitter(aes(x=arid.gradient, y=ihs(biomass), fill=Microsite, color=Microsite), size=3)+
# scale_color_manual(values=c("#E69F00","#56B4E9"))+ stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=ihs(biomass)), color="#56B4E9", data=subset(mean.spp, Microsite=="shrub")) + stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=ihs(biomass)), data=subset(mean.spp, Microsite=="open"), color="#E69F00")+ ylab("Annual biomass") + xlab("aridity gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16))
# RDA community and environmental data
Phytometer analysis
Difference in effect size
#
# library(bootES)
#
# effect.cal <- function(x){ bootES(x, R=999, data.col="Biomass", group.col="Microsite", contrast=c(shrub=1,open=-1), effect.type=c("cohens.d"))}
#
# effect.site <- by(subset(comm2016, Site != "Barstow"), comm2016$Site[comm2016$Site != "Barstow"], FUN=effect.cal)
#
# site.summary <- rbind(summary(effect.site$Panoche),summary(effect.site$Cuyama),summary(effect.site$TejonRanch),data.frame(stat=0,ci.low=0,ci.high=0,bias=0,std.error=0),summary(effect.site$MojavePreserve),summary(effect.site$SheepholeValley),summary(effect.site$Tecopa))
# site.summary[5,] <- 0
# site.summary <- sapply(site.summary, as.numeric)
#
# par(mar=c(4.5,4.5,.5,.5))
# plot(c(1:7),site.summary[,"stat"], xlim=c(0.5,7.5), ylim=c(-1,3), pch=19, cex=1.4, ylab="Hedge's G", xlab="", xaxt="n")
# arrows(c(1:7),as.numeric(site.summary[,"ci.high"]),c(1:7),site.summary[,"ci.low"], angle=90, code=3, length=0, lwd=2)
# abline(h=0, lty=2, lwd=2)
# axis(1, c(1:7), labels=site.names, cex.axis=1)
#
# ## phytometer
#
# effect.cal <- function(x){ bootES(x, R=999, data.col="phyto.biomass", group.col="Microsite", contrast=c(shrub=1,open=-1), effect.type=c("cohens.d"))}
#
# census2016 <- subset(census, Census=="end" & Year ==2016)
#
# effect.site <- by(subset(census2016, Site != "Barstow"), census2016$Site[census2016$Site != "Barstow"], FUN=effect.cal)
#
# site.summary <- rbind(summary(effect.site$Panoche),summary(effect.site$Cuyama),summary(effect.site$TejonRanch),data.frame(stat=0,ci.low=0,ci.high=0,bias=0,std.error=0),summary(effect.site$MojavePreserve),summary(effect.site$SheepholeValley),summary(effect.site$Tecopa))
# site.summary <- sapply(site.summary, as.numeric)
#
# par(mar=c(4.5,4.5,.5,.5))
# plot(c(1:7),site.summary[,"stat"], xlim=c(0.5,7.5), ylim=c(-1,3), pch=19, cex=1.4, ylab="Hedge's G", xlab="", xaxt="n")
# arrows(c(1:7),site.summary[,"ci.high"],c(1:7),site.summary[,"ci.low"], angle=90, code=3, length=0, lwd=2)
# abline(h=0, lty=2, lwd=2)
# axis(1, c(1:7), labels=site.names, cex.axis=1)Rii between years
Native non-native comparisons
## load species list
spp.list <- read.csv("Data/ERG.specieslist.csv")
## convert data to long format
status <- gather(community, species, abundance, 13:53)
names(spp.list)[1] <- "species" ## rename species column for merge
## combine native-non-native status with data set
status <- merge(status, spp.list, by="species")
## summarize per plot native and non-natives, convert back to wide format
mean.status <- status %>% group_by(Year, Site, Microsite, Rep, status) %>% summarize(abundance=sum(abundance))
mean.status <- spread(mean.status, status, abundance)
status.data <- mean.status
## calculate site means for native and non-native
mean.status <- mean.status %>% group_by(Year, Site, Microsite) %>% summarize(native=mean(native), non.native=mean(non.native))
## merge with original community data
mean.spp <- merge(mean.spp, mean.status, by=c("Year","Site","Microsite"))
## native
m1 <- lm(ihs(native) ~ poly(arid.gradient,2), data=subset(mean.spp, Microsite=="shrub"))
summary(m1)##
## Call:
## lm(formula = ihs(native) ~ poly(arid.gradient, 2), data = subset(mean.spp,
## Microsite == "shrub"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.03332 -0.65782 -0.04555 0.60465 1.25478
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.6533 0.2185 7.565 1.11e-05 ***
## poly(arid.gradient, 2)1 0.3447 0.8177 0.422 0.68145
## poly(arid.gradient, 2)2 -2.9468 0.8177 -3.604 0.00414 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.8177 on 11 degrees of freedom
## Multiple R-squared: 0.5448, Adjusted R-squared: 0.4621
## F-statistic: 6.583 on 2 and 11 DF, p-value: 0.01318m1 <- lm(ihs(native) ~ arid.gradient, data=subset(mean.spp, Microsite=="open"))
summary(m1)##
## Call:
## lm(formula = ihs(native) ~ arid.gradient, data = subset(mean.spp,
## Microsite == "open"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.89497 -0.73527 0.02181 0.50785 1.81716
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.0097 1.0111 3.966 0.00187 **
## arid.gradient 0.6575 0.3294 1.996 0.06913 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.093 on 12 degrees of freedom
## Multiple R-squared: 0.2493, Adjusted R-squared: 0.1867
## F-statistic: 3.984 on 1 and 12 DF, p-value: 0.06913m1 <- lm(ihs(native) ~ poly(arid.gradient,2)*Microsite, data=mean.spp)
summary(m1)##
## Call:
## lm(formula = ihs(native) ~ poly(arid.gradient, 2) * Microsite,
## data = mean.spp)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.6483 -0.6647 -0.1444 0.5383 2.0863
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) 2.0776 0.2616 7.943
## poly(arid.gradient, 2)1 3.0858 1.3841 2.229
## poly(arid.gradient, 2)2 -1.1145 1.3841 -0.805
## Micrositeshrub -0.4243 0.3699 -1.147
## poly(arid.gradient, 2)1:Micrositeshrub -2.5983 1.9574 -1.327
## poly(arid.gradient, 2)2:Micrositeshrub -3.0529 1.9574 -1.560
## Pr(>|t|)
## (Intercept) 6.65e-08 ***
## poly(arid.gradient, 2)1 0.0363 *
## poly(arid.gradient, 2)2 0.4293
## Micrositeshrub 0.2637
## poly(arid.gradient, 2)1:Micrositeshrub 0.1980
## poly(arid.gradient, 2)2:Micrositeshrub 0.1331
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.9787 on 22 degrees of freedom
## Multiple R-squared: 0.4229, Adjusted R-squared: 0.2918
## F-statistic: 3.225 on 5 and 22 DF, p-value: 0.0247ggplot(mean.spp) + geom_jitter(aes(x=arid.gradient, y=ihs(native), fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=ihs(native)), color="black")+ ylab("Native abundance") + xlab("aridity gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
## non-native
m1 <- lm(ihs(non.native) ~ poly(arid.gradient,2), data=subset(mean.spp, Microsite=="shrub"))
summary(m1)##
## Call:
## lm(formula = ihs(non.native) ~ poly(arid.gradient, 2), data = subset(mean.spp,
## Microsite == "shrub"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.32477 -0.20348 -0.05303 0.53297 2.27555
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.0777 0.3340 9.214 1.66e-06 ***
## poly(arid.gradient, 2)1 6.5874 1.2498 5.271 0.000264 ***
## poly(arid.gradient, 2)2 0.7721 1.2498 0.618 0.549292
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.25 on 11 degrees of freedom
## Multiple R-squared: 0.7191, Adjusted R-squared: 0.668
## F-statistic: 14.08 on 2 and 11 DF, p-value: 0.0009268m1 <- lm(ihs(non.native) ~ poly(arid.gradient,2), data=subset(mean.spp, Microsite=="open"))
summary(m1)##
## Call:
## lm(formula = ihs(non.native) ~ poly(arid.gradient, 2), data = subset(mean.spp,
## Microsite == "open"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.66301 -0.60211 0.01432 0.55858 1.91804
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.8844 0.2828 10.200 6.06e-07 ***
## poly(arid.gradient, 2)1 6.1497 1.0581 5.812 0.000117 ***
## poly(arid.gradient, 2)2 0.4615 1.0581 0.436 0.671141
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.058 on 11 degrees of freedom
## Multiple R-squared: 0.7554, Adjusted R-squared: 0.7109
## F-statistic: 16.98 on 2 and 11 DF, p-value: 0.0004331m1 <- lm(ihs(non.native) ~ poly(arid.gradient,2)*Microsite, data=mean.spp)
summary(m1)##
## Call:
## lm(formula = ihs(non.native) ~ poly(arid.gradient, 2) * Microsite,
## data = mean.spp)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.32477 -0.43001 -0.01113 0.58589 2.27555
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) 2.8844 0.3095 9.320
## poly(arid.gradient, 2)1 8.6970 1.6376 5.311
## poly(arid.gradient, 2)2 0.6527 1.6376 0.399
## Micrositeshrub 0.1933 0.4377 0.442
## poly(arid.gradient, 2)1:Micrositeshrub 0.6190 2.3159 0.267
## poly(arid.gradient, 2)2:Micrositeshrub 0.4393 2.3159 0.190
## Pr(>|t|)
## (Intercept) 4.28e-09 ***
## poly(arid.gradient, 2)1 2.49e-05 ***
## poly(arid.gradient, 2)2 0.694
## Micrositeshrub 0.663
## poly(arid.gradient, 2)1:Micrositeshrub 0.792
## poly(arid.gradient, 2)2:Micrositeshrub 0.851
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.158 on 22 degrees of freedom
## Multiple R-squared: 0.7361, Adjusted R-squared: 0.6761
## F-statistic: 12.27 on 5 and 22 DF, p-value: 9.2e-06m1 <- lm(ihs(non.native) ~ arid.gradient*Microsite, data=mean.spp)
summary(m1)##
## Call:
## lm(formula = ihs(non.native) ~ arid.gradient * Microsite, data = mean.spp)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.5300 -0.4742 0.1206 0.7847 2.0747
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 8.3299 1.0394 8.014 3.06e-08 ***
## arid.gradient 1.8530 0.3386 5.472 1.26e-05 ***
## Micrositeshrub 0.5808 1.4699 0.395 0.696
## arid.gradient:Micrositeshrub 0.1319 0.4789 0.275 0.785
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.124 on 24 degrees of freedom
## Multiple R-squared: 0.7289, Adjusted R-squared: 0.695
## F-statistic: 21.51 on 3 and 24 DF, p-value: 5.506e-07ggplot(mean.spp) + geom_jitter(aes(x=arid.gradient, y=ihs(non.native), fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+ stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=ihs(non.native)), color="black")+ ylab("Non-native abundance") + xlab("aridity gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
## mixed models
m1 <- lmer(ihs(non.native) ~ arid.gradient*Microsite + (1|Year), data=mean.spp)
shapiro.test(resid(m1))##
## Shapiro-Wilk normality test
##
## data: resid(m1)
## W = 0.96481, p-value = 0.4502anova(m1)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## arid.gradient 81.053 81.053 1 23.956 73.255 9.498e-09 ***
## Microsite 0.197 0.197 1 23.000 0.178 0.6768
## arid.gradient:Microsite 0.096 0.096 1 23.000 0.087 0.7712
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m1) ## R squared values ## R2m R2c
## 0.7152612 0.7820962m2 <- lmer(ihs(native) ~ poly(arid.gradient,2)*Microsite + (1|Year), data=mean.spp)
shapiro.test(resid(m2))##
## Shapiro-Wilk normality test
##
## data: resid(m2)
## W = 0.97665, p-value = 0.7641anova(m2)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value
## poly(arid.gradient, 2) 10.1669 5.0834 2 22 5.3071
## Microsite 1.2604 1.2604 1 22 1.3159
## poly(arid.gradient, 2):Microsite 4.0179 2.0089 2 22 2.0973
## Pr(>F)
## poly(arid.gradient, 2) 0.01316 *
## Microsite 0.26366
## poly(arid.gradient, 2):Microsite 0.14666
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values ## R2m R2c
## 0.3739082 0.3739082Phylogenetics
library(picante)
library(ape)
library(brranching)
library(taxize)
## load species list to create tree
spp.list <- read.csv("Data/ERG.specieslist.csv")
## create a combined species name column and drop erinoginum because not to species level
taxa <- paste(spp.list$Genus, spp.list$Species.name)
taxa <- taxa[taxa!="Erinoginum spp"]
## create tree of phylogeny
tree <- phylomatic(taxa=taxa, get = 'POST')
## calculate branch lengths using Grafen's method
## Grafen, A. (1989) The phylogenetic regression. Philosophical Transactions of the Royal society of London. Series B. Biological Sciences, 326, 119–157.
tree2 <- compute.brlen(tree, method="Grafen")
## view tree and outputs to verify branches
plot(tree2)
tree2##
## Phylogenetic tree with 39 tips and 29 internal nodes.
##
## Tip labels:
## cryptantha_barbigera, cryptantha_intermedia, phacelia_crenulata, phacelia_tanacetifolia, pectocarya_spp, amsinckia_grandiflora, ...
## Node labels:
## poales_to_asterales, eudicots, , , asterids, ericales_to_asterales, ...
##
## Rooted; includes branch lengths.## format community to only have site and microsite
comm <- community %>% group_by(Year, Site, Microsite) %>% summarise_if(is.numeric, funs(sum(., na.rm=T)))
comm <- comm[,c(1:3,13:53)] ## extract species abundances only
comm <- data.frame(comm) ## convert to dataframe
comm[,"site.micro"] <- paste(comm$Site,comm$Microsite,as.character(comm$Year)) ## create site by microsite column
rownames(comm) <- comm[,length(comm)] ## replace row names with site by microsite
comm[,c("Year","Site","Microsite","Erinoginum.spp","Boraginaceae.sp.","site.micro")] <- NULL ## drop all columns but ones for analysis
## match species name format
spp.list[,"spp.name"] <- paste(spp.list$Genus, spp.list$Species.name)
colnames(comm) <- spp.list$spp.name[match(colnames(comm), spp.list$Species.shorthand)]
## match species format
names(comm) <- gsub(" ", "_", names(comm), fixed=T) ## underscores instead of spaces
names(comm) <- gsub("__", "_", names(comm), fixed=T) ## replace double underscores with one
names(comm) <- tolower(names(comm)) ## lower case names
## mean phylogenetic distance
mpd.data <- mpd(comm, cophenetic(tree2), abundance.weighted=T)
## format dataframe
mpd.data <- data.frame(Year=c(rep(2016,14),rep(2017,14)),Microsite=c("open","shrub"),site=rownames(comm), mpd=mpd.data)
mpd.data[,"Site"] <- gsub(" open", "", mpd.data[,"site"])
mpd.data[,"Site"] <- gsub(" shrub", "", mpd.data[,"Site"])
mpd.data[,"Site"] <- gsub(" 2016", "", mpd.data[,"Site"])
mpd.data[,"Site"] <- gsub(" 2017", "", mpd.data[,"Site"])
mpd.data[,"site"] <- NULL
mpd.data[is.na(mpd.data)] <- 0 ## barstow no plants
## getspecies richness for sites
spp.rich <- community %>% group_by(Year, Site, Microsite) %>% summarize(richness=mean(Richness))
spp.rich <- data.frame(spp.rich)
## join data
mpd.rich <- merge(mpd.data, spp.rich, by=c("Site","Microsite","Year"))
mpd.arid <- merge(mpd.rich, mean.spp, by=c("Site","Microsite","Year"))
## phylogenetic distance
mpd.arid <- mpd.arid[c(2,4:nrow(mpd.arid)),] ## drop barstow
m1 <- lm(mpd~arid.gradient*Microsite, data=mpd.arid)
summary(m1)##
## Call:
## lm(formula = mpd ~ arid.gradient * Microsite, data = mpd.arid)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.51050 -0.19590 0.06299 0.14972 0.71607
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.97618 0.30321 3.219 0.00395 **
## arid.gradient 0.04716 0.10394 0.454 0.65451
## Micrositeshrub -1.07891 0.42881 -2.516 0.01966 *
## arid.gradient:Micrositeshrub -0.30441 0.14700 -2.071 0.05031 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2938 on 22 degrees of freedom
## Multiple R-squared: 0.3146, Adjusted R-squared: 0.2211
## F-statistic: 3.365 on 3 and 22 DF, p-value: 0.03692ggplot(mpd.arid) + geom_jitter(aes(x=arid.gradient, y=mpd, fill=Microsite, color=Microsite), size=3) +
scale_color_manual(values=c("#E69F00","#56B4E9")) + stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=mpd), data=subset(mpd.arid, Microsite=="shrub"), color="#56B4E9")+ stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=mpd), data=subset(mpd.arid, Microsite=="open"), color="#E69F00")+ ylab("Phylogenetic Community Dissimilarity") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m1 <- lm(mpd~arid.gradient, data=subset(mpd.arid, Microsite=="shrub"))
summary(m1)##
## Call:
## lm(formula = mpd ~ arid.gradient, data = subset(mpd.arid, Microsite ==
## "shrub"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.51050 -0.23112 0.03664 0.14204 0.71607
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.1027 0.3563 -0.288 0.778
## arid.gradient -0.2573 0.1222 -2.106 0.059 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3452 on 11 degrees of freedom
## Multiple R-squared: 0.2873, Adjusted R-squared: 0.2225
## F-statistic: 4.435 on 1 and 11 DF, p-value: 0.05899m2 <- lm(mpd~arid.gradient, data=subset(mpd.arid, Microsite=="open"))
summary(m2)##
## Call:
## lm(formula = mpd ~ arid.gradient, data = subset(mpd.arid, Microsite ==
## "open"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.38078 -0.09025 0.08952 0.15668 0.33783
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.97618 0.23852 4.093 0.00178 **
## arid.gradient 0.04716 0.08177 0.577 0.57575
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2311 on 11 degrees of freedom
## Multiple R-squared: 0.02935, Adjusted R-squared: -0.05889
## F-statistic: 0.3326 on 1 and 11 DF, p-value: 0.5758Changes in Mechanisms along gradient
### Nutrients
nutrient.mean <- nutrients.climate %>% group_by(Site, arid.gradient, microsite) %>% summarize(nitrogen=mean(N), phosphorus=mean(P), potassium=mean(K))
nutrient.mean <- data.frame(nutrient.mean)
## nitrogen
m1 <- lm(nitrogen ~ arid.gradient, data=subset(nutrient.mean, microsite=="open"))
summary(m1)##
## Call:
## lm(formula = nitrogen ~ arid.gradient, data = subset(nutrient.mean,
## microsite == "open"))
##
## Residuals:
## 1 3 5 7 9 11 13
## 1.62742 0.10479 -1.35216 0.97441 -0.05162 -0.76608 -0.53677
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2.7141 1.5041 -1.804 0.1310
## arid.gradient -1.2879 0.4313 -2.986 0.0306 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.124 on 5 degrees of freedom
## Multiple R-squared: 0.6407, Adjusted R-squared: 0.5689
## F-statistic: 8.917 on 1 and 5 DF, p-value: 0.03058anova(m1)## Analysis of Variance Table
##
## Response: nitrogen
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 11.263 11.263 8.9174 0.03058 *
## Residuals 5 6.315 1.263
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1ggplot(nutrient.mean) + geom_jitter(aes(x=arid.gradient, y=nitrogen, fill=microsite, color=microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9")) +
stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=nitrogen), color="#56B4E9", data=subset(nutrient.mean, microsite=="shrub")) +
stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=nitrogen), color="#E69F00", data=subset(nutrient.mean, microsite=="open"))+ ylab("Nitrogen") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
## phosphorus
m2 <- lm(phosphorus ~ arid.gradient * microsite, data=nutrient.mean)
summary(m2)##
## Call:
## lm(formula = phosphorus ~ arid.gradient * microsite, data = nutrient.mean)
##
## Residuals:
## Min 1Q Median 3Q Max
## -6.1684 -2.3472 0.6103 1.9735 7.0340
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 8.8933 5.3802 1.653 0.129
## arid.gradient 0.7555 1.5428 0.490 0.635
## micrositeshrub 10.5952 7.6088 1.392 0.194
## arid.gradient:micrositeshrub 2.0483 2.1819 0.939 0.370
##
## Residual standard error: 4.02 on 10 degrees of freedom
## Multiple R-squared: 0.3967, Adjusted R-squared: 0.2158
## F-statistic: 2.192 on 3 and 10 DF, p-value: 0.152anova(m2)## Analysis of Variance Table
##
## Response: phosphorus
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 43.011 43.011 2.6614 0.1339
## microsite 1 49.031 49.031 3.0339 0.1122
## arid.gradient:microsite 1 14.243 14.243 0.8813 0.3700
## Residuals 10 161.610 16.161ggplot(nutrient.mean) + geom_jitter(aes(x=arid.gradient, y=phosphorus, fill=microsite, color=microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9")) ## not significant
m3 <- lm(potassium ~ arid.gradient * microsite, data=nutrient.mean)
summary(m3)##
## Call:
## lm(formula = potassium ~ arid.gradient * microsite, data = nutrient.mean)
##
## Residuals:
## Min 1Q Median 3Q Max
## -160.67 -64.64 -14.22 64.63 206.92
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 290.78 145.57 1.998 0.0737 .
## arid.gradient 44.64 41.74 1.069 0.3101
## micrositeshrub 278.33 205.87 1.352 0.2062
## arid.gradient:micrositeshrub 56.24 59.03 0.953 0.3633
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 108.8 on 10 degrees of freedom
## Multiple R-squared: 0.4843, Adjusted R-squared: 0.3295
## F-statistic: 3.13 on 3 and 10 DF, p-value: 0.07444anova(m3)## Analysis of Variance Table
##
## Response: potassium
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 71879 71879 6.0754 0.0334 *
## microsite 1 28476 28476 2.4069 0.1519
## arid.gradient:microsite 1 10736 10736 0.9074 0.3633
## Residuals 10 118313 11831
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## potassium
ggplot(nutrient.mean) + geom_jitter(aes(x=arid.gradient, y=potassium, fill=microsite, color=microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9")) +
stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=potassium), color="black") 
#### Shrubs increase nitrogen availability at aridity extremes #### Shrubs do not effect potassium or phosphorus availability
se <- function(x) sd(na.omit(x))/sqrt(length(na.omit(x)))
smc <- census %>% group_by(Year, Site, Microsite, Census) %>% summarize(swc.avg=mean(swc), swc.se=se(swc))
smc.arid <- merge(smc, site.climate, by=c("Site","Year"))
## early swc
smc.early <- subset(smc.arid, Census=="emergence")
ggplot(smc.early) + geom_jitter(aes(x=arid.gradient, y=swc.avg, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=swc.avg), color="black") + ylab("Percent soil Moisture Content (emergence)") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m1 <- lm(swc.avg~ poly(arid.gradient,2) * Microsite, data=smc.early)
summary(m1)##
## Call:
## lm(formula = swc.avg ~ poly(arid.gradient, 2) * Microsite, data = smc.early)
##
## Residuals:
## Min 1Q Median 3Q Max
## -8.8989 -2.2340 0.1864 2.2093 8.7293
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) 11.1650 1.3550 8.240
## poly(arid.gradient, 2)1 32.2475 7.1700 4.498
## poly(arid.gradient, 2)2 20.8675 7.1700 2.910
## Micrositeshrub -0.4679 1.9163 -0.244
## poly(arid.gradient, 2)1:Micrositeshrub -3.3904 10.1399 -0.334
## poly(arid.gradient, 2)2:Micrositeshrub -3.1002 10.1399 -0.306
## Pr(>|t|)
## (Intercept) 3.6e-08 ***
## poly(arid.gradient, 2)1 0.000179 ***
## poly(arid.gradient, 2)2 0.008109 **
## Micrositeshrub 0.809376
## poly(arid.gradient, 2)1:Micrositeshrub 0.741276
## poly(arid.gradient, 2)2:Micrositeshrub 0.762672
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 5.07 on 22 degrees of freedom
## Multiple R-squared: 0.699, Adjusted R-squared: 0.6306
## F-statistic: 10.22 on 5 and 22 DF, p-value: 3.647e-05## end swc
smc.end <- subset(smc.arid, Census=="end")
ggplot(smc.end) + geom_jitter(aes(x=arid.gradient, y=swc.avg, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=swc.avg), color="black") + ylab("Percent soil Moisture Content (end)") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m1 <- lm(swc.avg~ poly(arid.gradient,2) * Microsite, data=smc.end)
summary(m1)##
## Call:
## lm(formula = swc.avg ~ poly(arid.gradient, 2) * Microsite, data = smc.end)
##
## Residuals:
## Min 1Q Median 3Q Max
## -8.7611 -2.4087 0.0009 2.0666 9.3760
##
## Coefficients:
## Estimate Std. Error t value
## (Intercept) 5.093 1.335 3.815
## poly(arid.gradient, 2)1 15.502 7.065 2.194
## poly(arid.gradient, 2)2 17.571 7.065 2.487
## Micrositeshrub -0.576 1.888 -0.305
## poly(arid.gradient, 2)1:Micrositeshrub -1.945 9.991 -0.195
## poly(arid.gradient, 2)2:Micrositeshrub -3.453 9.991 -0.346
## Pr(>|t|)
## (Intercept) 0.000946 ***
## poly(arid.gradient, 2)1 0.039073 *
## poly(arid.gradient, 2)2 0.020948 *
## Micrositeshrub 0.763211
## poly(arid.gradient, 2)1:Micrositeshrub 0.847426
## poly(arid.gradient, 2)2:Micrositeshrub 0.732914
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4.996 on 22 degrees of freedom
## Multiple R-squared: 0.4604, Adjusted R-squared: 0.3377
## F-statistic: 3.754 on 5 and 22 DF, p-value: 0.01312Shrubs do not alter soil moisture availability
hobo.means <- HOBO.data %>% group_by(season, Microsite, Site) %>% summarize(temp.avg=mean(Temp),temp.var=var(Temp),temp.se=se(Temp),rh.avg=mean(RH, na.rm=T),rh.var=var(RH, na.rm=T),rh.se=se(RH),frost=(sum(na.omit(Temp)<0)/length(na.omit(Temp))*100),heat=(sum(na.omit(Temp)>30)/length(na.omit(Temp))*100))
hobo.means[,"Year"] <- c(rep(2016,14),rep(2017,14))
hobo.arid <- merge(hobo.means, site.climate, by=c("Site","Year"))
## temperature
ggplot(hobo.arid) + geom_jitter(aes(x=arid.gradient, y=temp.avg, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~poly(x,2),aes(x=arid.gradient, y=temp.avg), color="black") + ylab("Average Temperature (C°)") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m1 <- lm(temp.avg~arid.gradient * Microsite, data=hobo.arid)
summary(m1)##
## Call:
## lm(formula = temp.avg ~ arid.gradient * Microsite, data = hobo.arid)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.2362 -0.7204 -0.3884 0.4197 3.2981
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.21950 1.38277 4.498 0.000149 ***
## arid.gradient -1.75909 0.45045 -3.905 0.000669 ***
## Micrositeshrub -1.42763 1.95554 -0.730 0.472428
## arid.gradient:Micrositeshrub 0.00911 0.63704 0.014 0.988709
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.495 on 24 degrees of freedom
## Multiple R-squared: 0.6064, Adjusted R-squared: 0.5571
## F-statistic: 12.32 on 3 and 24 DF, p-value: 4.451e-05anova(m1)## Analysis of Variance Table
##
## Response: temp.avg
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 67.815 67.815 30.3425 1.154e-05 ***
## Microsite 1 14.807 14.807 6.6252 0.01666 *
## arid.gradient:Microsite 1 0.000 0.000 0.0002 0.98871
## Residuals 24 53.639 2.235
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m1.s <- lm(temp.avg~arid.gradient, data=subset(hobo.arid, Microsite=="shrub"))
summary(m1.s)##
## Call:
## lm(formula = temp.avg ~ arid.gradient, data = subset(hobo.arid,
## Microsite == "shrub"))
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.5836 -0.9458 -0.4008 0.4122 3.2981
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.7919 1.3727 3.491 0.00446 **
## arid.gradient -1.7500 0.4472 -3.913 0.00206 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.484 on 12 degrees of freedom
## Multiple R-squared: 0.5607, Adjusted R-squared: 0.5241
## F-statistic: 15.31 on 1 and 12 DF, p-value: 0.00206## RH
ggplot(hobo.arid) + geom_jitter(aes(x=arid.gradient, y=rh.avg, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=rh.avg), color="black") + ylab("Relative Humidity (%)") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m2 <- lm(log(rh.avg)~arid.gradient * Microsite, data=hobo.arid)
summary(m2)##
## Call:
## lm(formula = log(rh.avg) ~ arid.gradient * Microsite, data = hobo.arid)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.313369 -0.094086 0.008993 0.111107 0.252701
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.703640 0.165465 28.427 < 2e-16 ***
## arid.gradient 0.251051 0.053902 4.658 9.92e-05 ***
## Micrositeshrub 0.049072 0.234003 0.210 0.836
## arid.gradient:Micrositeshrub 0.004987 0.076229 0.065 0.948
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.1789 on 24 degrees of freedom
## Multiple R-squared: 0.6497, Adjusted R-squared: 0.6059
## F-statistic: 14.84 on 3 and 24 DF, p-value: 1.131e-05anova(m2)## Analysis of Variance Table
##
## Response: log(rh.avg)
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 1.41614 1.41614 44.2510 7.014e-07 ***
## Microsite 1 0.00829 0.00829 0.2591 0.6154
## arid.gradient:Microsite 1 0.00014 0.00014 0.0043 0.9484
## Residuals 24 0.76806 0.03200
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## frost
hobo.arid2 <- subset(hobo.arid, Site!="Cuyama") ## drop cuyama because abnormal frost periods
ggplot(hobo.arid2) + geom_jitter(aes(x=arid.gradient, y=frost, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~x,aes(x=arid.gradient, y=frost), color="black") + ylab("Number of frost days") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m3 <- lm(frost~arid.gradient * Microsite, data=hobo.arid2)
summary(m3)##
## Call:
## lm(formula = frost ~ arid.gradient * Microsite, data = hobo.arid2)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.2695 -1.3394 -0.3817 1.9954 3.3744
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10.7896 2.2613 4.771 0.000116 ***
## arid.gradient 1.4999 0.7049 2.128 0.045970 *
## Micrositeshrub -2.3006 3.1980 -0.719 0.480218
## arid.gradient:Micrositeshrub -0.3756 0.9969 -0.377 0.710319
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.117 on 20 degrees of freedom
## Multiple R-squared: 0.3059, Adjusted R-squared: 0.2018
## F-statistic: 2.938 on 3 and 20 DF, p-value: 0.05818anova(m3)## Analysis of Variance Table
##
## Response: frost
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 31.066 31.0655 6.9304 0.01596 *
## Microsite 1 7.806 7.8058 1.7414 0.20187
## arid.gradient:Microsite 1 0.636 0.6363 0.1419 0.71032
## Residuals 20 89.649 4.4825
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## temperature variability
ggplot(hobo.arid) + geom_jitter(aes(x=arid.gradient, y=temp.var.x, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=temp.var.x), color="#E69F00", data=subset(hobo.arid, Microsite=="open")) + ylab("Number of frost days")+
stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=temp.var.x), color="#56B4E9", data=subset(hobo.arid, Microsite=="shrub")) + ylab("temperature variation") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m4 <- lm(log(temp.var.x)~arid.gradient * Microsite, data=hobo.arid)
summary(m4)##
## Call:
## lm(formula = log(temp.var.x) ~ arid.gradient * Microsite, data = hobo.arid)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.75540 -0.15564 0.02433 0.17182 0.46013
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 3.83222 0.26061 14.705 1.67e-13 ***
## arid.gradient -0.19133 0.08490 -2.254 0.0336 *
## Micrositeshrub -0.46416 0.36856 -1.259 0.2200
## arid.gradient:Micrositeshrub -0.02535 0.12006 -0.211 0.8346
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2818 on 24 degrees of freedom
## Multiple R-squared: 0.51, Adjusted R-squared: 0.4488
## F-statistic: 8.327 on 3 and 24 DF, p-value: 0.0005707anova(m4)## Analysis of Variance Table
##
## Response: log(temp.var.x)
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 0.91677 0.91677 11.5480 0.002368 **
## Microsite 1 1.06294 1.06294 13.3892 0.001241 **
## arid.gradient:Microsite 1 0.00354 0.00354 0.0446 0.834588
## Residuals 24 1.90531 0.07939
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## extreme heat
## reverse inverse hyperbolic sine
hs <- function(x) {
y <- 0.5*exp(-x)*(exp(2*x)-1)
return(y)
}
ggplot(hobo.arid) + geom_jitter(aes(x=arid.gradient, y=heat, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+
stat_smooth(method="lm", formula= y~hs(x),aes(x=arid.gradient, y=heat), color="#E69F00", data=subset(hobo.arid, Microsite=="open")) +
stat_smooth(method="lm", formula= y~hs(x),aes(x=arid.gradient, y=heat), color="#56B4E9", data=subset(hobo.arid, Microsite=="shrub")) + ylab("extreme heat days variation") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m5 <- lm(ihs(heat)~arid.gradient * Microsite, data=hobo.arid)
summary(m5)##
## Call:
## lm(formula = ihs(heat) ~ arid.gradient * Microsite, data = hobo.arid)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.7532 -0.3832 -0.1102 0.2072 1.2432
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.2199 0.4899 -0.449 0.657489
## arid.gradient -0.7396 0.1596 -4.635 0.000105 ***
## Micrositeshrub -0.1480 0.6928 -0.214 0.832615
## arid.gradient:Micrositeshrub 0.1787 0.2257 0.792 0.436163
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5296 on 24 degrees of freedom
## Multiple R-squared: 0.6529, Adjusted R-squared: 0.6095
## F-statistic: 15.05 on 3 and 24 DF, p-value: 1.015e-05anova(m5)## Analysis of Variance Table
##
## Response: ihs(heat)
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 9.3144 9.3144 33.2088 6.124e-06 ***
## Microsite 1 3.1726 3.1726 11.3114 0.00258 **
## arid.gradient:Microsite 1 0.1759 0.1759 0.6271 0.43616
## Residuals 24 6.7315 0.2805
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1shrubs reduce temperature variability
shrubz <- read.csv("Data/ERG.shrub.csv")
shrub.mean <- shrubz %>% group_by(Site, Microsite) %>% summarize(compact=mean(Compaction))
shrub.clim <- merge(shrub.mean, site.climate, by=c("Site"))
#shrub.clim <- subset(shrub.clim, Year==2016)
ggplot(shrub.clim) + geom_jitter(aes(x=arid.gradient, y=compact, fill=Microsite, color=Microsite), size=3)+
scale_color_manual(values=c("#E69F00","#56B4E9"))+ stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=compact), color="#E69F00", data=subset(shrub.clim, Microsite=="open"))+
stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=compact), color="#56B4E9", data=subset(shrub.clim, Microsite=="shrub")) + ylab("soil compaction") + xlab("Aridity Gradient")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),panel.background = element_blank(),axis.text=element_text(size=14),axis.title=element_text(size=16)) 
m1 <- lm(log(compact) ~ arid.gradient* Microsite, data=shrub.clim)
summary(m1)##
## Call:
## lm(formula = log(compact) ~ arid.gradient * Microsite, data = shrub.clim)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.36122 -0.13659 -0.02823 0.05479 0.61368
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.26645 0.23301 1.144 0.26410
## arid.gradient -0.07328 0.07590 -0.965 0.34394
## Micrositeshrub -1.19983 0.32952 -3.641 0.00130 **
## arid.gradient:Micrositeshrub -0.32933 0.10734 -3.068 0.00528 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2519 on 24 degrees of freedom
## Multiple R-squared: 0.5932, Adjusted R-squared: 0.5424
## F-statistic: 11.67 on 3 and 24 DF, p-value: 6.53e-05anova(m1)## Analysis of Variance Table
##
## Response: log(compact)
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 1.24730 1.24730 19.6548 0.0001755 ***
## Microsite 1 0.37672 0.37672 5.9362 0.0226248 *
## arid.gradient:Microsite 1 0.59733 0.59733 9.4127 0.0052769 **
## Residuals 24 1.52305 0.06346
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1shrubs reduce soil compaction and increase substrate penetration
Phytometer models
library(lsmeans)## The 'lsmeans' package is being deprecated.
## Users are encouraged to switch to 'emmeans'.
## See help('transition') for more information, including how
## to convert 'lsmeans' objects and scripts to work with 'emmeans'.##
## Attaching package: 'lsmeans'## The following object is masked from 'package:lmerTest':
##
## lsmeanscensus[is.na(census)] <- 0
census[,"phyto.abd"] <- rowSums(census[,c("Phacelia","Plantago","Salvia")])
census.end <- subset(census, Census=="end")
census.end[,"Year"] <- as.factor(census.end$Year)
census.arid <- merge(census.end, site.climate, by=c("Year","Site"))
## all species end of season
m1 <- glm.nb(phyto.abd~ Microsite * Nutrient * arid.gradient, data=census.arid)
anova(m2, test="Chisq") ## Site and microsite*nutrient significant## Analysis of Variance Table
##
## Response: log(rh.avg)
## Df Sum Sq Mean Sq F value Pr(>F)
## arid.gradient 1 1.41614 1.41614 44.2510 7.014e-07 ***
## Microsite 1 0.00829 0.00829 0.2591 0.6154
## arid.gradient:Microsite 1 0.00014 0.00014 0.0043 0.9484
## Residuals 24 0.76806 0.03200
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## run full model for phacelia biomass
m1 <- aov(log(Phacelia.biomass)~ Microsite * Nutrient * arid.gradient, data=subset(census.arid, Phacelia.biomass>0))
summary(m1) ## site and microsite## Df Sum Sq Mean Sq F value Pr(>F)
## Microsite 1 19.0 19.00 9.126 0.00281 **
## Nutrient 1 8.8 8.76 4.207 0.04141 *
## arid.gradient 1 122.7 122.66 58.927 4.95e-13 ***
## Microsite:Nutrient 1 0.2 0.16 0.078 0.78078
## Microsite:arid.gradient 1 1.4 1.41 0.678 0.41121
## Nutrient:arid.gradient 1 0.0 0.00 0.000 0.99128
## Microsite:Nutrient:arid.gradient 1 2.3 2.27 1.088 0.29795
## Residuals 225 468.3 2.08
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1TukeyHSD(m1, "Microsite")## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Nutrient, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, Nutrient, arid.gradient## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(Phacelia.biomass) ~ Microsite * Nutrient * arid.gradient, data = subset(census.arid, Phacelia.biomass > 0))
##
## $Microsite
## diff lwr upr p adj
## shrub-open 0.5995862 0.2084783 0.9906941 0.0028111TukeyHSD(m1, "Nutrient")## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Nutrient, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, Nutrient, arid.gradient## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(Phacelia.biomass) ~ Microsite * Nutrient * arid.gradient, data = subset(census.arid, Phacelia.biomass > 0))
##
## $Nutrient
## diff lwr upr p adj
## Ambient-Added -0.3865601 -0.7590709 -0.01404918 0.0420304## run full model for plantago biomass
m2 <- aov(log(Plantago.biomass)~ Microsite * Nutrient * arid.gradient, data=subset(census.arid, Plantago.biomass>0))
summary(m2) ## site and year + sitexmicrosite## Df Sum Sq Mean Sq F value Pr(>F)
## Microsite 1 9.59 9.59 5.718 0.018067 *
## Nutrient 1 26.14 26.14 15.582 0.000122 ***
## arid.gradient 1 136.39 136.39 81.293 1.01e-15 ***
## Microsite:Nutrient 1 0.61 0.61 0.366 0.546337
## Microsite:arid.gradient 1 1.57 1.57 0.937 0.334540
## Nutrient:arid.gradient 1 3.96 3.96 2.362 0.126491
## Microsite:Nutrient:arid.gradient 1 3.21 3.21 1.912 0.168873
## Residuals 146 244.96 1.68
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1TukeyHSD(m2, "Microsite")## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Nutrient, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, Nutrient, arid.gradient## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(Plantago.biomass) ~ Microsite * Nutrient * arid.gradient, data = subset(census.arid, Plantago.biomass > 0))
##
## $Microsite
## diff lwr upr p adj
## shrub-open -0.5018942 -0.9167114 -0.08707699 0.0180668TukeyHSD(m2, "Nutrient")## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Nutrient, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, Nutrient, arid.gradient## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(Plantago.biomass) ~ Microsite * Nutrient * arid.gradient, data = subset(census.arid, Plantago.biomass > 0))
##
## $Nutrient
## diff lwr upr p adj
## Ambient-Added -0.8205043 -1.23339 -0.4076185 0.0001319## run full model for salvia biomass
m3 <- aov(log(Salvia.biomass)~ Microsite * Nutrient * arid.gradient, data=subset(census.arid, Salvia.biomass>0))
summary(m3) ## site and year + sitexmicrosite## Df Sum Sq Mean Sq F value Pr(>F)
## Microsite 1 0.1 0.11 0.054 0.815771
## Nutrient 1 24.8 24.84 12.494 0.000465 ***
## arid.gradient 1 33.7 33.73 16.965 4.79e-05 ***
## Microsite:Nutrient 1 0.2 0.23 0.115 0.734281
## Microsite:arid.gradient 1 0.6 0.59 0.299 0.585029
## Nutrient:arid.gradient 1 20.3 20.28 10.199 0.001538 **
## Microsite:Nutrient:arid.gradient 1 0.0 0.00 0.000 0.988929
## Residuals 338 672.1 1.99
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1TukeyHSD(m3, "Microsite")## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Nutrient, arid.gradient## Warning in replications(paste("~", xx), data = mf): non-factors ignored:
## Microsite, Nutrient, arid.gradient## Tukey multiple comparisons of means
## 95% family-wise confidence level
##
## Fit: aov(formula = log(Salvia.biomass) ~ Microsite * Nutrient * arid.gradient, data = subset(census.arid, Salvia.biomass > 0))
##
## $Microsite
## diff lwr upr p adj
## shrub-open 0.03535378 -0.26289 0.3335975 0.8157715m3.n <- lm(log(Salvia.biomass)~ arid.gradient, data=subset(census.arid, Salvia.biomass>0 & Nutrient =="Added"))
m3.a <- lm(log(Salvia.biomass)~ arid.gradient, data=subset(census.arid, Salvia.biomass>0 & Nutrient =="Ambient"))
## abundance
## run full model for phacelia abundance
m1 <- glm.nb(Phacelia~ Microsite * Nutrient * arid.gradient, data=census.arid)
summary(m1) ## site and microsite##
## Call:
## glm.nb(formula = Phacelia ~ Microsite * Nutrient * arid.gradient,
## data = census.arid, init.theta = 0.2308296117, link = log)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.03830 -0.81531 -0.76103 0.01823 3.03660
##
## Coefficients:
## Estimate Std. Error z value
## (Intercept) -0.0007756 0.5883363 -0.001
## Micrositeshrub 1.2705913 0.8115605 1.566
## NutrientAmbient 0.1077404 0.8333728 0.129
## arid.gradient 0.1858087 0.1944620 0.956
## Micrositeshrub:NutrientAmbient 0.1849576 1.1409892 0.162
## Micrositeshrub:arid.gradient 0.2621592 0.2696848 0.972
## NutrientAmbient:arid.gradient 0.0421013 0.2759568 0.153
## Micrositeshrub:NutrientAmbient:arid.gradient -0.0505405 0.3789622 -0.133
## Pr(>|z|)
## (Intercept) 0.999
## Micrositeshrub 0.117
## NutrientAmbient 0.897
## arid.gradient 0.339
## Micrositeshrub:NutrientAmbient 0.871
## Micrositeshrub:arid.gradient 0.331
## NutrientAmbient:arid.gradient 0.879
## Micrositeshrub:NutrientAmbient:arid.gradient 0.894
##
## (Dispersion parameter for Negative Binomial(0.2308) family taken to be 1)
##
## Null deviance: 585.87 on 839 degrees of freedom
## Residual deviance: 550.93 on 832 degrees of freedom
## AIC: 1941.5
##
## Number of Fisher Scoring iterations: 1
##
##
## Theta: 0.2308
## Std. Err.: 0.0231
##
## 2 x log-likelihood: -1923.4540anova(m1, test="Chisq")## Warning in anova.negbin(m1, test = "Chisq"): tests made without re-
## estimating 'theta'## Analysis of Deviance Table
##
## Model: Negative Binomial(0.2308), link: log
##
## Response: Phacelia
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev
## NULL 839 585.87
## Microsite 1 19.9145 838 565.96
## Nutrient 1 1.0729 837 564.89
## arid.gradient 1 11.4485 836 553.44
## Microsite:Nutrient 1 1.0254 835 552.41
## Microsite:arid.gradient 1 1.4563 834 550.96
## Nutrient:arid.gradient 1 0.0064 833 550.95
## Microsite:Nutrient:arid.gradient 1 0.0167 832 550.93
## Pr(>Chi)
## NULL
## Microsite 8.098e-06 ***
## Nutrient 0.3002852
## arid.gradient 0.0007155 ***
## Microsite:Nutrient 0.3112354
## Microsite:arid.gradient 0.2275227
## Nutrient:arid.gradient 0.9362121
## Microsite:Nutrient:arid.gradient 0.8971935
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1lsmeans(m1, pairwise~"Microsite")## NOTE: Results may be misleading due to involvement in interactions## $lsmeans
## Microsite lsmean SE df asymp.LCL asymp.UCL
## open -0.5548206 0.1205597 NA -0.7911133 -0.3185280
## shrub 0.1120837 0.1124787 NA -0.1083705 0.3325379
##
## Results are averaged over the levels of: Nutrient
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df z.ratio p.value
## open - shrub -0.6669043 0.1648821 NA -4.045 0.0001
##
## Results are averaged over the levels of: Nutrient
## Results are given on the log (not the response) scale.## run full model for plantago biomass
m2 <- glm.nb(Plantago~ Microsite * Nutrient * arid.gradient, data=census.arid)
summary(m2) ## site and microsite##
## Call:
## glm.nb(formula = Plantago ~ Microsite * Nutrient * arid.gradient,
## data = census.arid, init.theta = 0.1287536289, link = log)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8053 -0.7211 -0.6578 -0.5689 2.4129
##
## Coefficients:
## Estimate Std. Error z value
## (Intercept) 1.0031515 0.7094622 1.414
## Micrositeshrub -2.8208986 1.0387678 -2.716
## NutrientAmbient -0.0810403 1.0160835 -0.080
## arid.gradient 0.3420451 0.2337751 1.463
## Micrositeshrub:NutrientAmbient 0.5580731 1.4672180 0.380
## Micrositeshrub:arid.gradient -0.7058890 0.3368271 -2.096
## NutrientAmbient:arid.gradient 0.0826214 0.3361349 0.246
## Micrositeshrub:NutrientAmbient:arid.gradient -0.0006522 0.4776644 -0.001
## Pr(>|z|)
## (Intercept) 0.15737
## Micrositeshrub 0.00662 **
## NutrientAmbient 0.93643
## arid.gradient 0.14343
## Micrositeshrub:NutrientAmbient 0.70368
## Micrositeshrub:arid.gradient 0.03611 *
## NutrientAmbient:arid.gradient 0.80584
## Micrositeshrub:NutrientAmbient:arid.gradient 0.99891
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for Negative Binomial(0.1288) family taken to be 1)
##
## Null deviance: 427.38 on 839 degrees of freedom
## Residual deviance: 414.78 on 832 degrees of freedom
## AIC: 1596.1
##
## Number of Fisher Scoring iterations: 1
##
##
## Theta: 0.1288
## Std. Err.: 0.0138
##
## 2 x log-likelihood: -1578.1100anova(m2, test="Chisq")## Warning in anova.negbin(m2, test = "Chisq"): tests made without re-
## estimating 'theta'## Analysis of Deviance Table
##
## Model: Negative Binomial(0.1288), link: log
##
## Response: Plantago
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev Pr(>Chi)
## NULL 839 427.38
## Microsite 1 5.5584 838 421.82 0.01839
## Nutrient 1 0.0296 837 421.79 0.86334
## arid.gradient 1 0.5610 836 421.23 0.45386
## Microsite:Nutrient 1 1.6621 835 419.56 0.19732
## Microsite:arid.gradient 1 4.7218 834 414.84 0.02978
## Nutrient:arid.gradient 1 0.0649 833 414.78 0.79889
## Microsite:Nutrient:arid.gradient 1 0.0000 832 414.78 0.99859
##
## NULL
## Microsite *
## Nutrient
## arid.gradient
## Microsite:Nutrient
## Microsite:arid.gradient *
## Nutrient:arid.gradient
## Microsite:Nutrient:arid.gradient
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1lsmeans(m2, pairwise~"Microsite")## NOTE: Results may be misleading due to involvement in interactions## $lsmeans
## Microsite lsmean SE df asymp.LCL asymp.UCL
## open -0.1639688 0.1466943 NA -0.4514843 0.1235467
## shrub -0.6304166 0.1522519 NA -0.9288248 -0.3320084
##
## Results are averaged over the levels of: Nutrient
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df z.ratio p.value
## open - shrub 0.4664478 0.2114234 NA 2.206 0.0274
##
## Results are averaged over the levels of: Nutrient
## Results are given on the log (not the response) scale.## run full model for salvia biomass
m3 <- glm.nb(Salvia~ Microsite * Nutrient * arid.gradient, data=census.arid)
summary(m3) ## site and microsite##
## Call:
## glm.nb(formula = Salvia ~ Microsite * Nutrient * arid.gradient,
## data = census.arid, init.theta = 0.3061627656, link = log)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.3585 -1.1226 -0.8020 0.2021 2.2024
##
## Coefficients:
## Estimate Std. Error z value
## (Intercept) 2.39712 0.46918 5.109
## Micrositeshrub -0.02997 0.67161 -0.045
## NutrientAmbient 0.61208 0.67236 0.910
## arid.gradient 0.57763 0.15655 3.690
## Micrositeshrub:NutrientAmbient 0.16062 0.94881 0.169
## Micrositeshrub:arid.gradient 0.07096 0.22516 0.315
## NutrientAmbient:arid.gradient 0.24763 0.22629 1.094
## Micrositeshrub:NutrientAmbient:arid.gradient -0.14470 0.31899 -0.454
## Pr(>|z|)
## (Intercept) 3.24e-07 ***
## Micrositeshrub 0.964406
## NutrientAmbient 0.362646
## arid.gradient 0.000225 ***
## Micrositeshrub:NutrientAmbient 0.865571
## Micrositeshrub:arid.gradient 0.752644
## NutrientAmbient:arid.gradient 0.273802
## Micrositeshrub:NutrientAmbient:arid.gradient 0.650108
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for Negative Binomial(0.3062) family taken to be 1)
##
## Null deviance: 785.98 on 839 degrees of freedom
## Residual deviance: 721.96 on 832 degrees of freedom
## AIC: 2994
##
## Number of Fisher Scoring iterations: 1
##
##
## Theta: 0.3062
## Std. Err.: 0.0230
##
## 2 x log-likelihood: -2976.0340anova(m3, test="Chisq")## Warning in anova.negbin(m3, test = "Chisq"): tests made without re-
## estimating 'theta'## Analysis of Deviance Table
##
## Model: Negative Binomial(0.3062), link: log
##
## Response: Salvia
##
## Terms added sequentially (first to last)
##
##
## Df Deviance Resid. Df Resid. Dev Pr(>Chi)
## NULL 839 785.98
## Microsite 1 0.258 838 785.72 0.61139
## Nutrient 1 3.609 837 782.11 0.05747
## arid.gradient 1 54.596 836 727.51 1.48e-13
## Microsite:Nutrient 1 4.529 835 722.98 0.03332
## Microsite:arid.gradient 1 0.005 834 722.98 0.94599
## Nutrient:arid.gradient 1 0.870 833 722.11 0.35089
## Microsite:Nutrient:arid.gradient 1 0.145 832 721.96 0.70327
##
## NULL
## Microsite
## Nutrient .
## arid.gradient ***
## Microsite:Nutrient *
## Microsite:arid.gradient
## Nutrient:arid.gradient
## Microsite:Nutrient:arid.gradient
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1lsmeans(m3, pairwise~Microsite*Nutrient)## NOTE: Results may be misleading due to involvement in interactions## $lsmeans
## Microsite Nutrient lsmean SE df asymp.LCL asymp.UCL
## open Added 0.6995864 0.1351555 NA 0.4346865 0.9644863
## shrub Added 0.4610742 0.1382722 NA 0.1900656 0.7320828
## open Ambient 0.5839160 0.1380287 NA 0.3133848 0.8544472
## shrub Ambient 0.9312613 0.1338667 NA 0.6688874 1.1936351
##
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df z.ratio p.value
## open,Added - shrub,Added 0.2385122 0.1933552 NA 1.234 0.6054
## open,Added - open,Ambient 0.1156704 0.1931811 NA 0.599 0.9325
## open,Added - shrub,Ambient -0.2316749 0.1902296 NA -1.218 0.6154
## shrub,Added - open,Ambient -0.1228418 0.1953743 NA -0.629 0.9228
## shrub,Added - shrub,Ambient -0.4701871 0.1924565 NA -2.443 0.0692
## open,Ambient - shrub,Ambient -0.3473453 0.1922816 NA -1.806 0.2702
##
## Results are given on the log (not the response) scale.
## P value adjustment: tukey method for comparing a family of 4 estimates## treatment plots
census.long <- gather(census.arid, species, biomass, 12:14)
census.long[,"species"] <- ifelse(census.long[,"species"] =="Phacelia.biomass", "P. tanacetifolia",
ifelse(census.long[,"species"]=="Plantago.biomass","P. insularis","S. columbariae"))
census.long[,"species"] <- as.factor(census.long$species)
## microsite average
census.micro <- census.long %>% filter(biomass>0) %>% group_by(Microsite, species) %>% summarize(Biomass=mean(biomass), error=se(biomass))
ggplot(census.micro, aes(x=species, y=Biomass, fill=Microsite)) + geom_bar(position=position_dodge(), stat="identity", colour="black") + scale_fill_brewer()+ geom_errorbar(aes(ymin=Biomass-error, ymax=Biomass+error),
width=.2, # Width of the error bars
position=position_dodge(.9))+ theme_Publication() + theme(axis.text.x = element_text(size=12, face="italic"))## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## nutrient average
census.nut <- census.long %>% filter(biomass>0) %>% group_by(Nutrient, species, Year) %>% summarize(Biomass=mean(biomass), error=se(biomass))
ggplot(census.nut, aes(x=species, y=Biomass, fill=Nutrient)) + geom_bar(position=position_dodge(), stat="identity",colour="black") + scale_fill_brewer()+ geom_errorbar(aes(ymin=Biomass-error, ymax=Biomass+error),
width=.2, # Width of the error bars
position=position_dodge(.9))+ theme_Publication()+ theme(axis.text.x = element_text(size=12, face="italic")) +facet_grid(~Year)## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## aridity average
ggplot(census.long, aes(x=arid.gradient, y=biomass, fill=species, colour=species)) + theme_Publication()+ stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=biomass), color="#009E73", fill="#009E7340", data=subset(census.long, species=="Phacelia.biomass" & biomass>0)) + stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=biomass), color="#E69F00",fill="#E69F0040", data=subset(census.long, species=="Plantago.biomass" & biomass>0))+ stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=biomass), color="#56B4E9", fill="#56B4E940", data=subset(census.long, species=="Salvia.biomass" & biomass>0))## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
### need to separate zeros using gamma-hurdle model
##http://seananderson.ca/2014/05/18/gamma-hurdle.html
## logistic first
census.long[,"presence"] <- ifelse(census.long$biomass>0,1,0)
m1.l <- glmer(presence~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year),data=subset(census.long, species=="P. tanacetifolia"), family=binomial)
car::Anova(m1.l)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: presence
## Chisq Df Pr(>Chisq)
## Microsite 28.8440 1 7.845e-08 ***
## poly(arid.gradient, 2) 13.7669 2 0.001025 **
## Nutrient 0.0065 1 0.935820
## Microsite:poly(arid.gradient, 2) 1.6297 2 0.442697
## Microsite:Nutrient 1.9145 1 0.166465
## poly(arid.gradient, 2):Nutrient 0.5605 2 0.755599
## Microsite:poly(arid.gradient, 2):Nutrient 1.5360 2 0.463938
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m1 <- lmer(log(biomass)~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year),data=subset(census.long, species=="P. tanacetifolia" & presence==1))
shapiro.test(resid(m1))##
## Shapiro-Wilk normality test
##
## data: resid(m1)
## W = 0.99429, p-value = 0.5244anova(m1)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF
## Microsite 10.828 10.828 1 220.60
## poly(arid.gradient, 2) 96.805 48.403 2 193.15
## Nutrient 7.785 7.785 1 220.10
## Microsite:poly(arid.gradient, 2) 3.723 1.862 2 220.61
## Microsite:Nutrient 0.068 0.068 1 220.08
## poly(arid.gradient, 2):Nutrient 4.093 2.046 2 220.02
## Microsite:poly(arid.gradient, 2):Nutrient 2.078 1.039 2 220.06
## F.value Pr(>F)
## Microsite 5.6947 0.01786 *
## poly(arid.gradient, 2) 25.4555 1.54e-10 ***
## Nutrient 4.0941 0.04424 *
## Microsite:poly(arid.gradient, 2) 0.9790 0.37731
## Microsite:Nutrient 0.0357 0.85037
## poly(arid.gradient, 2):Nutrient 1.0762 0.34269
## Microsite:poly(arid.gradient, 2):Nutrient 0.5464 0.57983
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m1) ## R squared values ## R2m R2c
## 0.2463743 0.3096945m2.l <- glmer(presence~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year),data=subset(census.long, species=="P. insularis"), family=binomial)
car::Anova(m2.l)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: presence
## Chisq Df Pr(>Chisq)
## Microsite 1.1386 1 0.285954
## poly(arid.gradient, 2) 24.7326 2 4.26e-06 ***
## Nutrient 0.3030 1 0.582001
## Microsite:poly(arid.gradient, 2) 9.8407 2 0.007296 **
## Microsite:Nutrient 3.7231 1 0.053664 .
## poly(arid.gradient, 2):Nutrient 9.3310 2 0.009415 **
## Microsite:poly(arid.gradient, 2):Nutrient 1.8086 2 0.404829
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m2 <- lmer(log(biomass)~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year), data=subset(census.long, species=="P. insularis" & presence==1))
shapiro.test(resid(m2))##
## Shapiro-Wilk normality test
##
## data: resid(m2)
## W = 0.9915, p-value = 0.49anova(m2)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF
## Microsite 0.4051 0.4051 1 141.10
## poly(arid.gradient, 2) 13.1478 6.5739 2 141.72
## Nutrient 7.8434 7.8434 1 141.28
## Microsite:poly(arid.gradient, 2) 0.4061 0.2031 2 141.30
## Microsite:Nutrient 0.4193 0.4193 1 141.23
## poly(arid.gradient, 2):Nutrient 1.6505 0.8252 2 141.24
## Microsite:poly(arid.gradient, 2):Nutrient 0.3606 0.1803 2 141.17
## F.value Pr(>F)
## Microsite 0.3818 0.537661
## poly(arid.gradient, 2) 6.1948 0.002635 **
## Nutrient 7.3911 0.007377 **
## Microsite:poly(arid.gradient, 2) 0.1914 0.826049
## Microsite:Nutrient 0.3951 0.530642
## poly(arid.gradient, 2):Nutrient 0.7777 0.461438
## Microsite:poly(arid.gradient, 2):Nutrient 0.1699 0.843906
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values ## R2m R2c
## 0.1291956 0.6996902m3.l <- glmer(presence~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year),data=subset(census.long, species=="S. columbariae"), family=binomial)
car::Anova(m3.l)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: presence
## Chisq Df Pr(>Chisq)
## Microsite 0.0208 1 0.88536
## poly(arid.gradient, 2) 72.4822 2 < 2e-16 ***
## Nutrient 0.9592 1 0.32739
## Microsite:poly(arid.gradient, 2) 2.2173 2 0.33000
## Microsite:Nutrient 3.8296 1 0.05036 .
## poly(arid.gradient, 2):Nutrient 2.1024 2 0.34951
## Microsite:poly(arid.gradient, 2):Nutrient 0.3469 2 0.84075
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m3 <- lmer(log(Salvia.biomass)~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year), data=subset(census.arid, Salvia.biomass>0))
shapiro.test(resid(m3))##
## Shapiro-Wilk normality test
##
## data: resid(m3)
## W = 0.99339, p-value = 0.1337anova(m3)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF
## Microsite 1.519 1.5187 1 333.08
## poly(arid.gradient, 2) 52.112 26.0561 2 333.65
## Nutrient 15.470 15.4701 1 333.09
## Microsite:poly(arid.gradient, 2) 1.015 0.5076 2 333.15
## Microsite:Nutrient 0.244 0.2444 1 333.03
## poly(arid.gradient, 2):Nutrient 17.550 8.7750 2 333.04
## Microsite:poly(arid.gradient, 2):Nutrient 4.472 2.2362 2 333.06
## F.value Pr(>F)
## Microsite 0.9019 0.342966
## poly(arid.gradient, 2) 15.4738 3.745e-07 ***
## Nutrient 9.1871 0.002628 **
## Microsite:poly(arid.gradient, 2) 0.3014 0.739951
## Microsite:Nutrient 0.1451 0.703473
## poly(arid.gradient, 2):Nutrient 5.2112 0.005909 **
## Microsite:poly(arid.gradient, 2):Nutrient 1.3280 0.266399
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m3) ## R squared values ## R2m R2c
## 0.1322110 0.2743503## set up prediction dataset
library(effects)## Loading required package: carData## Use the command
## lattice::trellis.par.set(effectsTheme())
## to customize lattice options for effects plots.
## See ?efffectTheme for details.epha <- as.data.frame(Effect(c("arid.gradient"),m1.l))
epla <- as.data.frame(Effect(c("arid.gradient"),m2.l))
esal <- as.data.frame(Effect(c("arid.gradient"),m3.l))
ee <- rbind(epha,epla, esal)
ee[,"species"] <- rep(unique(census.long$species),each=5)
ggplot(data=ee, aes(arid.gradient,fit,color=species))+geom_line(lwd=1.4) +xlab("aridity gradient") + ylab("probability of occurrence") + theme_Publication() + geom_ribbon(aes(ymin=lower, ymax=upper, fill=species), alpha=0.5)## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Calculate effects size estimates for regressions
library(effects)
epha <- as.data.frame(Effect(c("arid.gradient"),m1))
epla <- as.data.frame(Effect(c("arid.gradient"),m2))
esal <- as.data.frame(Effect(c("arid.gradient"),m3))
ee <- rbind(epha,epla, esal)
ee[,"species"] <- rep(unique(census.long$species),each=5)
#ee[,2:5] <- exp(ee[,2:5])
## add colours for species
color.add<- subset(census.long, biomass>0)
color.add$color <- ifelse(color.add$species=="P. tanacetifolia", "#009E73",ifelse(color.add$species=="P. insularis", "#D55E00","#0072B2"))
ggplot()+ geom_jitter(aes(x=arid.gradient, y=log(biomass), group=species), color=color.add$color,alpha=0.4, width = 0.05, data=subset(census.long, biomass>0)) +geom_line(data=ee, aes(arid.gradient,fit,color=species),lwd=1.4) +xlab("aridity gradient") + ylab("log-transformed biomass") + theme_Publication()## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
#
# ## to check for overdispersion
# overdisp_fun <- function(model) {
# ## number of variance parameters in
# ## an n-by-n variance-covariance matrix
# vpars <- function(m) {
# nrow(m)*(nrow(m)+1)/2
# }
# model.df <- sum(sapply(VarCorr(model),vpars))+length(fixef(model))
# (rdf <- nrow(model@frame)-model.df)
# rp <- residuals(model)
# Pearson.chisq <- sum(rp^2)
# prat <- Pearson.chisq/rdf
# pval <- pchisq(Pearson.chisq, df=rdf, lower.tail=FALSE,log.p=TRUE)
# c(chisq=Pearson.chisq,ratio=prat,p=exp(pval))
# }
## related packages
# library(car)
# library(lme4)
## https://www.rdocumentation.org/packages/lme4/versions/1.1-15/topics/pvalues
## https://arxiv.org/pdf/1406.5823.pdf
## Calculating Pvalues from glmer
# ##Provides parametric bootstrapping methods for model comparisons
# library(afex)
# mixed(Phacelia~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year), data=census.arid, family=poisson, control=glmerControl(optimizer="bobyqa",optCtrl=list(maxfun=100000)), method="LRT")
# ## accounting for overdispersion in GLMM
# ## Harrison, X. A. (2014). Using observation-level random effects to model overdispersion in count data in ecology and evolution. PeerJ, 2, e616.
# m4 <- glmer(Phacelia~Microsite * arid.gradient * Nutrient + (1|Year), data=census.arid, family=poisson)
# #car::Anova(m4, test.statistics="LR", type=c("III"))
# overdisp_fun(m4) ## over dispersion?
#
#
# ## re-run with observational-level random effects
# disp<-seq(nrow(census.arid))
# m4 <- glmer(Phacelia~Microsite * arid.gradient * Nutrient + (1|Year) + (1|disp), data=census.arid, family=poisson)
# car::Anova(m4, test.statistics="LR", type=c("III"))
#
# m4 <- glmer.nb(Phacelia~Microsite * arid.gradient * Nutrient + (1|Year), data=census.arid)
# car::Anova(m4, test.statistics="LR", type=c("III"))
#
# m5 <- glmer(Plantago~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year), data=census.arid, family=poisson)
# overdisp_fun(m5) ## over dispersion?
#
# ## re-run with observational-level random effects
# m5 <- glmer(Plantago~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year)+ (1|disp), data=census.arid, family=poisson)
# car::Anova(m5, test.statistics="LR", type=c("III"))
#
# m6 <- glmer(Salvia~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year), data=census.arid, family=poisson)
# overdisp_fun(m6) ## over dispersion?
#
#
# ## re-run with observational-level random effects
# m6 <- glmer(Salvia~Microsite * poly(arid.gradient,2) * Nutrient + (1|Year)+ (1|disp), data=census.arid, family=poisson)
# car::Anova(m6, test.statistics="LR", type=c("III"))Community analysis
## Regressions To Separate Phylogenetic Attraction And Repulsion
## sppregs(community[,13:53], )
### Network analysis
library(igraph)
library(statnet)
## reload native vs non-native species
status <- read.csv("Data//ERG.specieslist.csv")
shrub.comm <- subset(community, Microsite=="shrub", select=13:53)
shrub.comm <- shrub.comm[,colSums(shrub.comm)!=0]
open.comm <- subset(community, Microsite=="open", select=13:53)
open.comm <- open.comm[,colSums(open.comm)!=0]
## Shrub
## correlation species network
network <- data.frame(cor(shrub.comm))
network[,"To"] <- colnames(network)
test <- gather(network, From, correlation, 1:(ncol(network)-1))
test <- subset(test, correlation>0.10 & correlation != 1 | correlation < -0.10 )
n.shrub <- graph_from_data_frame(test[,1:2], directed=TRUE)
## plot network circle
lo <- layout_in_circle(n.shrub)
plot(n.shrub, layout=lo)
transitivity(n.shrub, type = "local")
transitivity(n.shrub, type = "global")
## colour graphs
shapes <- V(n.shrub)$name %in% (subset(status, status=="non.native", select=1))[,1] +1
plot(n.shrub, layout=layout_on_grid(n.shrub),vertex.shape=c("circle", "square")[shapes], vertex.color=c("tomato2", "royalblue")[shapes])
## open
## correlation species network
network <- data.frame(cor(open.comm))
network[,"To"] <- colnames(network)
test <- gather(network, From, correlation, 1:(ncol(network)-1))
test <- subset(test, correlation>0.10 & correlation != 1 | correlation < -0.10 )
n.open <- graph_from_data_frame(test[,1:2], directed=TRUE)
## plot network circle
lo <- layout_in_circle(n.open)
plot(n.open, layout=lo)
## colour graphs
shapes <- V(n.open)$name %in% (subset(status, status=="non.native", select=1))[,1] +1
plot(n.open, layout=layout_on_grid(n.open),vertex.shape=c("circle", "square")[shapes], vertex.color=c("tomato2", "royalblue")[shapes])
## statistics
transitivity(n.open, type = "local")
transitivity(n.open, type = "global") indicator species analysis
library(indicspecies)
## indicator species analysis
indval <- multipatt(community[,13:53], community$Microsite, control=how(nperm=999))
summary(indval)##
## Multilevel pattern analysis
## ---------------------------
##
## Association function: IndVal.g
## Significance level (alpha): 0.05
##
## Total number of species: 41
## Selected number of species: 14
## Number of species associated to 1 group: 14
##
## List of species associated to each combination:
##
## Group open #sps. 8
## stat p.value
## E.cicutarium 0.558 0.001 ***
## L.gracilis 0.427 0.001 ***
## A.wrangelianus 0.201 0.001 ***
## C.brevipes 0.171 0.029 *
## A.lentiginosus 0.149 0.030 *
## C.rigida 0.149 0.016 *
## A.grandiflora 0.129 0.014 *
## Boraginaceae.sp. 0.120 0.018 *
##
## Group shrub #sps. 6
## stat p.value
## C.fremontii 0.379 0.001 ***
## M.glabrata 0.277 0.001 ***
## P.tanacetifolia 0.267 0.001 ***
## A.tessellata 0.199 0.002 **
## B.nigra 0.179 0.012 *
## E.nitidum 0.172 0.008 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1site.names <- as.character(unique(community$Site))
community.site <- subset(community, Site==site.names[7])
## species accumilation curves
accum1 <- specaccum(subset(community.site, Microsite=="shrub", select=13:53), method="random", permutations=1000)
accum2 <- specaccum(subset(community.site, Microsite=="open", select=13:53), method="random", permutations=1000)
par(mar=c(4.5,4.5,0.5,0.5))
plot(accum2, col="#E69F00",ci.col="#E69F0090", xlab="Sampling effort", cex.axis=1.5, cex.lab=1.8, ylab="Species richness", lwd=2, ylim=c(0,16))
plot(accum1,add=T , col="#56B4E9", ci.col="#56B4E990", lwd=2)
micros <- community$Microsite[1:120]
indval <- multipatt(community.site[,13:53], micros, control=how(nperm=999))
summary(indval)##
## Multilevel pattern analysis
## ---------------------------
##
## Association function: IndVal.g
## Significance level (alpha): 0.05
##
## Total number of species: 41
## Selected number of species: 4
## Number of species associated to 1 group: 4
##
## List of species associated to each combination:
##
## Group open #sps. 1
## stat p.value
## Boraginaceae.sp. 0.316 0.025 *
##
## Group shrub #sps. 3
## stat p.value
## A.tessellata 0.531 0.001 ***
## B.nigra 0.475 0.012 *
## M.affinis 0.343 0.046 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1cluster analysis
library("cluster")
library("factoextra")
library("magrittr")
comm.sum <- community %>% group_by(Microsite, Site) %>% summarize_if(is.numeric, funs(sum))
comm.sum <- data.frame(comm.sum)
rownames(comm.sum) <- paste(comm.sum$Microsite,comm.sum$Site)
## coreelation matrix
### Distance measures
res.dist <- get_dist(comm.sum[,13:53], stand = TRUE, method = "euclidean")
fviz_dist(res.dist, gradient = list(low = "#00AFBB", mid = "white", high = "#FC4E07"))
## dendrogram
## transform data
comm.trans <- decostand(comm.sum[,13:53], "hell")
## clean data for ordination
## see distribution of spp
boxplot(comm.trans, xaxt="n")
labs <- colnames(comm.trans)
text(cex=0.8, x=1:41-1, y=-0.12, labs, xpd=TRUE, srt=45)
## remove spp with only one instancve
comm.trans <- comm.trans[,!colSums(comm.trans)==apply(comm.trans, 2, max)]
## replace outliers with mean
avg.max <- function(x) {
y = max(x)
avg = mean(x)
x[x==y] <- avg
return(x)
}
## dot chart to find outliers (remove or Avg)
#dotchart(comm.trans$E.cicutarium) ## Erodium
comm.trans[,"E.cicutarium"] <- avg.max(comm.trans$E.cicutarium) ## removed outlier
#dotchart(comm.trans$A.wrangelianus) ## Chilean trefoil
comm.trans[,"A.wrangelianus"] <- NULL ## removed entire species
#dotchart(comm.trans$A.lentiginosus) ##
comm.trans[,"A.lentiginosus"] <- avg.max(comm.trans$A.lentiginosus) ## removed outlier
#dotchart(comm.trans$H.vulgare) ## Hordeum
comm.trans[,"H.vulgare"] <- NULL ## removed entire species
#dotchart(comm.trans$Pectocarya.spp) ##
comm.trans[,"Pectocarya.spp"] <- avg.max(comm.trans$Pectocarya.spp) ## removed outlier
#dotchart(comm.trans$L.arizonicus) ## Lupin
comm.trans[,"L.arizonicus"] <- NULL ## removed entire species
#dotchart(comm.trans$M.bellioides) ##
comm.trans[,"M.bellioides"] <- NULL ## removed entire species
#dotchart(comm.trans$B.nigra) ## Mustard
comm.trans[,"B.nigra"] <- NULL ## removed entire species
#dotchart(comm.trans$N.demissum) ## Sand Matt
comm.trans[,"N.demissum"] <- avg.max(comm.trans$N.demissum) ## removed outlier
#dotchart(comm.trans$L.gracilis) ##
comm.trans[,"L.gracilis"] <- avg.max(comm.trans$L.gracilis) ## removed outlier
#dotchart(comm.trans$A.grandiflora) ## Amsinckia
comm.trans[,"A.grandiflora"] <- NULL ## removed entire species
#dotchart(comm.trans$B.diandrus) ## Bromus diandrus
comm.trans[,"B.diandrus"] <- NULL ## removed entire species
#dotchart(comm.trans$M.affinis) ##
comm.trans[,"M.affinis"] <- NULL ## removed entire species
#dotchart(comm.trans$P.crenulata) ##
comm.trans[,"P.crenulata"] <- NULL ## removed entire species
#dotchart(comm.trans$Erinoginum.spp) # buckwheat
comm.trans[,"Erinoginum.spp"] <- avg.max(comm.trans$Erinoginum.spp) ## removed outlier
#dotchart(comm.trans$C.lasiophyllus) #
comm.trans[,"C.lasiophyllus"] <- NULL ## removed entire species
library(usdm)
library(corrplot)
## Check for collinearity
vifcor(comm.trans)## 4 variables from the 25 input variables have collinearity problem:
##
## E.wallacei P.insularis E.glyptosperma L.decorus
##
## After excluding the collinear variables, the linear correlation coefficients ranges between:
## min correlation ( L.gracilis ~ S.barbatus ): -0.003266766
## max correlation ( C.fremontii ~ M.glabrata ): 0.8450794
##
## ---------- VIFs of the remained variables --------
## Variables VIF
## 1 B.rubens Inf
## 2 E.cicutarium Inf
## 3 S.barbatus Inf
## 4 C.barbigera Inf
## 5 C.claviformis Inf
## 6 A.lentiginosus Inf
## 7 C.rigida Inf
## 8 P.tanacetifolia Inf
## 9 Pectocarya.spp Inf
## 10 D.capitatum Inf
## 11 C.brevipes Inf
## 12 M.glabrata Inf
## 13 C.fremontii Inf
## 14 C.intermedia Inf
## 15 N.demissum Inf
## 16 L.gracilis Inf
## 17 A.tessellata Inf
## 18 P.secunda Inf
## 19 L.sparsiflorus Inf
## 20 Erinoginum.spp Inf
## 21 E.nitidum Inf## drop Claviformis because of collinear problems in P.insularis
comm.trans[,"C.claviformis"] <- NULL ## removed entire species
comm.trans[,"L.decorus"] <- NULL ## removed entire species
comm.trans[,"E.glyptosperma"] <- NULL ## removed entire species
comm.trans[,"E.wallacei"] <- NULL ## removed entire species
cor.dat <- cor(comm.trans[,1:20])
corrplot(cor.dat, method="number")
## calculate distances
dis <- vegdist(comm.trans, method="bray")
## cluster analysis
clus <- hclust(dis, "ward.D2")
## cluster colours
colz <- c("#3e4444", "#86af49", "#405d27","#c1946a")
fviz_dend(clus, k = 4, # Cut in four groups
cex = 0.7, cex.axis=2, # label size
k_colors = colz,
color_labels_by_k = TRUE, # color labels by groups
rect = TRUE # Add rectangle around groups
) 
grp <- cutree(clus, 4)
## Interpretation of classes
## get soil moisture from site/microsite
smc.avg <- smc.end %>% group_by(Site, Microsite) %>% summarize(mean.smc=mean(swc.avg, na.rm=T))
smc.avg <- data.frame(smc.avg)
smc.avg[,"grp"] <- as.factor(c(3,3,2,1,4,4,1,1,4,4,3,3,2,2))
boxplot(mean.smc ~ grp, data=smc.avg)
## CA or PCA
dca1 <- decorana(comm.trans) ## length of gradient >2 & determine relative differences in community composition
## conduct ordination
ord <- cca(comm.trans)
summary(ord)##
## Call:
## cca(X = comm.trans)
##
## Partitioning of mean squared contingency coefficient:
## Inertia Proportion
## Total 1.53 1
## Unconstrained 1.53 1
##
## Eigenvalues, and their contribution to the mean squared contingency coefficient
##
## Importance of components:
## CA1 CA2 CA3 CA4 CA5 CA6 CA7
## Eigenvalue 0.6615 0.2873 0.1656 0.10566 0.08765 0.07595 0.05451
## Proportion Explained 0.4324 0.1878 0.1082 0.06907 0.05729 0.04964 0.03563
## Cumulative Proportion 0.4324 0.6202 0.7284 0.79749 0.85478 0.90443 0.94006
## CA8 CA9 CA10 CA11 CA12 CA13
## Eigenvalue 0.03070 0.02546 0.01821 0.008974 0.00457 0.003798
## Proportion Explained 0.02006 0.01664 0.01190 0.005870 0.00299 0.002480
## Cumulative Proportion 0.96012 0.97676 0.98866 0.994530 0.99752 1.000000
##
## Scaling 2 for species and site scores
## * Species are scaled proportional to eigenvalues
## * Sites are unscaled: weighted dispersion equal on all dimensions
##
##
## Species scores
##
## CA1 CA2 CA3 CA4 CA5 CA6
## B.rubens -1.5272 0.37620 0.03693 -0.17971 0.26195 -0.007680
## E.cicutarium -0.5874 -0.19213 -0.20338 -0.20206 -0.36108 0.524976
## S.barbatus 0.1974 -0.64533 0.21879 0.05816 0.09674 -0.053334
## C.barbigera 0.4900 -0.11013 0.11027 0.09640 0.08716 0.152044
## A.lentiginosus 0.2199 -0.28948 -0.59265 -0.08500 0.77156 -0.296433
## C.rigida 0.4907 -0.59656 -0.53938 0.07713 0.37931 -0.042119
## P.tanacetifolia -1.2902 0.91318 -0.45764 2.53829 0.03756 -0.152588
## Pectocarya.spp 0.5495 -0.11384 -0.59153 -0.11011 0.04276 0.166931
## D.capitatum -0.7173 -0.38053 -0.34301 0.92059 -1.47031 1.938948
## C.brevipes 0.4996 -0.08694 -1.06985 -0.06884 -0.31678 0.438663
## P.insularis 0.8031 0.56576 -0.62483 -0.12389 0.05743 -0.096057
## M.glabrata 0.7884 0.84798 0.94548 -0.09754 -0.18919 0.009159
## C.fremontii 0.8264 0.71519 0.19154 -0.07910 -0.11619 0.016657
## C.intermedia 0.3301 -0.17327 -0.89534 -0.13610 -0.03923 -0.580017
## N.demissum 0.7716 0.44620 -0.72001 0.09033 0.21708 -0.241934
## L.gracilis -0.9869 -0.31033 0.07421 0.35879 -1.02622 -0.476879
## A.tessellata -1.0035 -0.36194 0.26967 -0.84774 -1.13864 -1.238336
## P.secunda -1.5685 0.52156 0.01749 0.80285 0.12080 -0.361127
## L.sparsiflorus 0.7841 1.12673 0.93614 0.24022 0.04660 0.069996
## Erinoginum.spp 0.5728 -1.00733 0.32438 0.36758 1.12435 -0.440374
## E.nitidum 0.8450 0.85547 -0.32831 -0.02364 -0.06814 -0.220852
##
##
## Site scores (weighted averages of species scores)
##
## CA1 CA2 CA3 CA4 CA5 CA6
## open Barstow 0.3623 -1.3253 0.59499 0.23046 0.56897 0.613534
## open Cuyama -0.2820 -0.9387 -0.39810 0.01254 -2.11537 2.774795
## open HeartofMojave 0.7431 0.2335 -2.05624 -0.52149 0.05849 0.033486
## open PanocheHills -1.6700 0.4724 0.08096 -1.23638 1.43764 0.474226
## open SheepholeValley 0.9186 0.9037 0.77143 0.17618 0.30513 0.116582
## open Tecopa 0.5210 -1.3909 -0.13250 0.41626 1.33826 -0.340390
## open TejonRanch -0.6455 -0.9594 0.23131 -0.09915 -1.38186 -1.124251
## shrub Barstow 0.5555 -0.5840 0.87451 0.02079 0.01283 0.724468
## shrub Cuyama -1.7597 0.9562 -0.21109 3.09525 0.07451 -0.062796
## shrub HeartofMojave 0.7511 0.7261 -1.52964 0.11670 -0.12309 -0.542992
## shrub PanocheHills -2.0371 0.9309 -0.05454 -1.74126 1.62903 1.240252
## shrub SheepholeValley 1.0489 1.6807 1.57641 -0.25781 -0.55721 -0.007623
## shrub Tecopa 0.4692 -1.4952 0.52997 0.40750 1.29037 -0.766954
## shrub TejonRanch -1.1319 -0.1862 0.35145 -1.12008 -1.13771 -1.730813## calculate priority
spp.priority <- colSums(comm.trans)
## plot ordination
par(mar=c(4.5,4.5,0.5,0.5))
plot(ord, type="n")
ordiellipse(ord, grp, lty = 2, col = "grey80", draw="polygon", alpha=150)
orditorp(ord, display = "species", cex = 0.7, col = "darkorange3", priority=spp.priority, air=0.5)
orditorp(ord, display = "sites", cex = 0.7, col = "darkslateblue", air=0.1)
##collect environmental variables for site
nutrients <- read.csv("Data/ERG.soilnutrients.csv")
nutrients.mean <- nutrients %>% group_by(microsite, Site) %>% summarise_all(funs(mean))
nutrients.vars <- data.frame(nutrients.mean)
## summarize daily means across sites
means <- HOBO.data %>% group_by(Microsite, Site) %>% summarize(Temp.=mean(Temp, na.rm=T),rh=mean(RH, na.rm=T),temp.se=se(Temp),rh.se=se(RH))
means <- data.frame(means)
## soil moisture
smc.avg <- smc.early %>% group_by(Microsite, Site) %>% summarize(SMC=mean(swc.avg, na.rm=T))
envs <- data.frame(swc=smc.avg[,"SMC"],nutrients.vars[,c("N","P","K")], means[,c("Temp.","rh")])
## Check for collinearity
cor(envs)## SMC N P K Temp. rh
## SMC 1.0000000 -0.31692385 0.44854712 0.81694831 -0.6306962 0.9176776
## N -0.3169238 1.00000000 0.05879919 -0.05275422 0.4188434 -0.4374109
## P 0.4485471 0.05879919 1.00000000 0.64885814 -0.5264257 0.5353719
## K 0.8169483 -0.05275422 0.64885814 1.00000000 -0.5705341 0.7740852
## Temp. -0.6306962 0.41884336 -0.52642569 -0.57053407 1.0000000 -0.8625519
## rh 0.9176776 -0.43741089 0.53537193 0.77408521 -0.8625519 1.0000000envs[,"K"] <- NULL ## drop potassium for being correlated
envs[,"rh"] <- NULL ## drop humidity for being correlated
ord.env <- envfit(ord, envs)
plot(ord.env)
# ordicluster(ord, clus, col="blue")
# ## conduct ordination with gradient analysis
# ord <- cca(comm.trans~SMC+N+P+Temp.,data=envs)
# (R2adj <- RsquareAdj(ord)$adj.r.squared) ## 30.0% of variation explained
#
# ord <- cca(comm.trans~.,data=envs)
# ord.0 <- cca(comm.trans~1,data=envs)
# m <- ordistep(ord.0, scope=formula(ord))
#
#
# ## plot ordination
# par(mar=c(4.5,4.5,0.5,0.5))
# plot(ord, type="n") # display="bp", ylim=c(-2,2),xlim=c(-2,2))
# ordiellipse(ord, grp, lty = 2, col = colz, draw="polygon", alpha=150)
# orditorp(ord, display = "species", cex = 0.7, col = "darkred", priority=spp.priority, air=0.5)
# orditorp(ord, display = "sites", cex = 0.7, col = "black", air=0.1)Short-term vs long-term effects
Repeated of analyses separating long- and short- term effects
Ordination including aridity and microsite (pCCA)
Partial Constrained Correspondence Analysis for community data
Mixed model resuilts
community.arid <- merge(community, gams.data, by=c("Year","Site"))
### Ambient
mean.year <- community.arid %>% group_by(Microsite, gams, Year) %>% summarize(abd=mean(Abundance),rich=mean(Richness), bio=mean(Biomass))
## combine other dataframes with gams
mpd.gams <- merge(mpd.arid, gams.data, by=c("Year","Site"))
mean.status2 <- merge(mean.spp,gams.data, by=c("Year","Site"))
mean.status2[,"native.ratio"] <- mean.status2$native/(mean.status2$native+mean.status2$non.native)
mean.status2[is.na(mean.status2)] <- 0
m1 <- lmer(ihs(bio)~Microsite * gams + (1|Year), data=mean.year)
shapiro.test(resid(m1))##
## Shapiro-Wilk normality test
##
## data: resid(m1)
## W = 0.9789, p-value = 0.8234anova(m1)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 1.6259 1.6259 1 23.000 11.782 0.002272 **
## gams 18.1519 18.1519 1 23.367 131.533 4.448e-11 ***
## Microsite:gams 1.0755 1.0755 1 23.000 7.794 0.010365 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m1) ## R squared values ## R2m R2c
## 0.549095 0.945669m2 <- lmer(ihs(abd)~Microsite * gams + (1|Year), data=mean.year)
shapiro.test(resid(m2))##
## Shapiro-Wilk normality test
##
## data: resid(m2)
## W = 0.97267, p-value = 0.6537car::Anova(m2, type=2)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: ihs(abd)
## Chisq Df Pr(>Chisq)
## Microsite 0.1875 1 0.6650
## gams 63.2057 1 1.862e-15 ***
## Microsite:gams 0.1083 1 0.7421
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values ## R2m R2c
## 0.6343625 0.8521801m2.nat <- lmer(ihs(native)~Microsite * poly(gams,2) + (1|Year), data=mean.status2)
shapiro.test(resid(m2.nat))##
## Shapiro-Wilk normality test
##
## data: resid(m2.nat)
## W = 0.98084, p-value = 0.8701anova(m2.nat)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 1.2604 1.2604 1 22 1.1237 0.30063
## poly(gams, 2) 8.0514 4.0257 2 22 3.5890 0.04477 *
## Microsite:poly(gams, 2) 2.5294 1.2647 2 22 1.1275 0.34184
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values ## R2m R2c
## 0.6343625 0.8521801m2.exo <- lmer(ihs(non.native)~Microsite * gams+ (1|Year), data=mean.status2)
shapiro.test(resid(m2.exo))##
## Shapiro-Wilk normality test
##
## data: resid(m2.exo)
## W = 0.90696, p-value = 0.01672anova(m2.exo)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 0.247 0.247 1 23.000 0.183 0.6727
## gams 74.389 74.389 1 23.965 55.056 1.175e-07 ***
## Microsite:gams 0.175 0.175 1 23.000 0.129 0.7223
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values ## R2m R2c
## 0.6343625 0.8521801m2.rat <- lmer(ihs(native.ratio)~Microsite *gams+ (1|Year), data=mean.status2)
shapiro.test(resid(m2.rat))##
## Shapiro-Wilk normality test
##
## data: resid(m2.rat)
## W = 0.89603, p-value = 0.009249anova(m2.rat)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 0.03367 0.03367 1 23.0000 0.3040 0.58669
## gams 0.47126 0.47126 1 9.7629 4.2555 0.06674 .
## Microsite:gams 0.02598 0.02598 1 23.0000 0.2346 0.63271
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values## R2m R2c
## 0.6343625 0.8521801m3 <- lmer(rich~Microsite * poly(gams,2) + (1|Year), data=mean.year)
shapiro.test(resid(m3))##
## Shapiro-Wilk normality test
##
## data: resid(m3)
## W = 0.93835, p-value = 0.1004car::Anova(m3, type=2)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: rich
## Chisq Df Pr(>Chisq)
## Microsite 0.6282 1 0.428012
## poly(gams, 2) 10.9223 2 0.004249 **
## Microsite:poly(gams, 2) 2.5197 2 0.283699
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m3) ## R squared values ## R2m R2c
## 0.3445392 0.3495475m4 <- lmer(mpd~Microsite * gams + (1|Year), data=mpd.gams)
shapiro.test(resid(m4))##
## Shapiro-Wilk normality test
##
## data: resid(m4)
## W = 0.98753, p-value = 0.9827anova(m4)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 0.68094 0.68094 1 22 9.2604 0.005967 **
## gams 0.27407 0.27407 1 22 3.7272 0.066522 .
## Microsite:gams 0.55289 0.55289 1 22 7.5189 0.011898 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m4) ## R squared values ## R2m R2c
## 0.3852134 0.3852134### Biomass
p1 <- ggplot(mean.year) + geom_jitter(aes(x=gams, y=ihs(bio)), color="#56B4E9", size=3, data=subset(mean.year, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=ihs(bio)), color="#E69F00", size=3, data=subset(mean.year, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~poly(x,3),aes(x=gams, y=ihs(bio)), color="#E69F00", fill="#E69F0080", data=subset(mean.year, Microsite=="open"), lwd=1.4) + stat_smooth(method="lm", formula= y~poly(x,3),aes(x=gams, y=ihs(bio)), color="#56B4E9", fill="#56B4E970", data=subset(mean.year, Microsite=="shrub"), lwd=1.4) + ylab("Annnual plant biomass")+ xlab("") #annotate("text", x=-4.5,y=4, label="a", size=8)+coord_cartesian(ylim=c(0, 4))
### abundance
p2 <- ggplot(mean.year) + geom_jitter(aes(x=gams, y=ihs(abd)), color="#56B4E9", size=3, data=subset(mean.year, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=ihs(abd)), color="#E69F00", size=3, data=subset(mean.year, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= ihs(y)~x,aes(x=gams, y=abd), color="black", fill="Grey50", data=mean.year, lwd=1.4)+ ylab("Annual plant abundance")+ xlab("")+ annotate("text", x=-4.5,y=6, label="b", size=8)
### richness
p3 <- ggplot(mean.year) + geom_jitter(aes(x=gams, y=rich), color="#56B4E9", size=3, data=subset(mean.year, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=rich), color="#E69F00", size=3, data=subset(mean.year, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=rich), color="#E69F00", fill="#E69F0080", data=subset(mean.year, Microsite=="open"), lwd=1.4) + stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=rich), color="#56B4E9", fill="#56B4E970", data=subset(mean.year, Microsite=="shrub"), lwd=1.4) + ylab("Species richness") + xlab("Aridity Gradient")+ annotate("text", x=-4.5,y=4, label="c", size=8)
### phylogenetic diveristy
p4 <- ggplot(mpd.gams)+ geom_jitter(aes(x=gams, y=mpd), color="#56B4E9", size=3, data=subset(mpd.gams, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=mpd), color="#E69F00", size=3, data=subset(mpd.gams, Microsite=="open")) + stat_smooth(method="lm", formula= y~x,aes(x=gams, y=mpd), data=subset(mpd.gams, Microsite=="shrub"), color="#56B4E9",fill="#56B4E980", lwd=1.4)+ stat_smooth(method="lm", formula= y~x,aes(x=gams, y=mpd), data=subset(mpd.gams, Microsite=="open"), color="#E69F00",fill="#E69F0080", lwd=1.4)+ ylab("Phylogenetic Community Dissimilarity") + xlab("Aridity Gradient")+ theme_Publication()+ annotate("text", x=-4.5,y=1.8, label="d", size=8)
grid.arrange(p1,p2,p3,p4)
### Nutrients
nutrient.mean <- nutrients.climate %>% group_by(Site, arid.gradient, microsite) %>% summarize(nitrogen=mean(N), phosphorus=mean(P), potassium=mean(K))
nutrient.mean <- data.frame(nutrient.mean,gams=rep(gams.data[gams.data$Year==2016,"gams"],each=2))
## potassium
m5 <- lm(potassium ~ poly(gams,2)* microsite, data=nutrient.mean)
summary(m5)##
## Call:
## lm(formula = potassium ~ poly(gams, 2) * microsite, data = nutrient.mean)
##
## Residuals:
## Min 1Q Median 3Q Max
## -39.224 -15.052 -5.456 7.166 51.901
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 141.46 12.13 11.662 2.67e-06 ***
## poly(gams, 2)1 -217.98 45.39 -4.803 0.001351 **
## poly(gams, 2)2 150.46 45.39 3.315 0.010617 *
## micrositeshrub 90.20 17.15 5.258 0.000766 ***
## poly(gams, 2)1:micrositeshrub -256.71 64.19 -4.000 0.003952 **
## poly(gams, 2)2:micrositeshrub 149.36 64.19 2.327 0.048389 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 32.09 on 8 degrees of freedom
## Multiple R-squared: 0.9641, Adjusted R-squared: 0.9416
## F-statistic: 42.95 on 5 and 8 DF, p-value: 1.438e-05anova(m5)## Analysis of Variance Table
##
## Response: potassium
## Df Sum Sq Mean Sq F value Pr(>F)
## poly(gams, 2) 2 170636 85318 82.835 4.502e-06 ***
## microsite 1 28476 28476 27.648 0.0007663 ***
## poly(gams, 2):microsite 2 22053 11026 10.706 0.0054741 **
## Residuals 8 8240 1030
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m5) ## R squared values ## R2m R2c
## 0.9429147 0.9429147p5 <- ggplot(nutrient.mean) + geom_jitter(aes(x=gams, y=potassium), color="#56B4E9", size=3, data=subset(nutrient.mean, microsite=="shrub"))+ geom_jitter(aes(x=gams, y=potassium), color="#E69F00", size=3, data=subset(nutrient.mean, microsite=="open"))+ stat_smooth(method="lm", formula= y~x,aes(x=gams, y=potassium), color="#56B4E9", fill="#56B4E980", data=subset(nutrient.mean, microsite=="shrub"), lwd=1.4) +
stat_smooth(method="lm", formula= y~x,aes(x=gams, y=potassium), color="#E69F00", fill="#E69F0080",data=subset(nutrient.mean, microsite=="open"), lwd=1.4)+ ylab("Nitrogen (ppm)") + xlab("")+ theme_Publication()+ annotate("text", x=-4.5,y=32, label="a", size=8)+coord_cartesian(ylim=c(0, 32))
se <- function(x) sd(na.omit(x))/sqrt(length(na.omit(x)))
smc <- census %>% group_by(Year, Site, Microsite, Census) %>% summarize(swc.avg=mean(swc), swc.se=se(swc))
smc.arid <- merge(smc, gams.data, by=c("Site","Year"))
## early swc
smc.early <- subset(smc.arid, Census=="emergence")
m6 <- lmer(swc.avg~Microsite * poly(gams,2) + (1|Year), data=smc.early)
shapiro.test(resid(m6))##
## Shapiro-Wilk normality test
##
## data: resid(m6)
## W = 0.91432, p-value = 0.02516anova(m6)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 1.53 1.53 1 21.000 0.084 0.7752
## poly(gams, 2) 1449.15 724.58 2 21.343 39.560 6.617e-08 ***
## Microsite:poly(gams, 2) 3.69 1.85 2 21.000 0.101 0.9046
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m6) ## R squared values ## R2m R2c
## 0.5464095 0.8953768### soil moisture at emergence
p6 <- ggplot(smc.early) + geom_jitter(aes(x=gams, y=swc.avg), color="#56B4E9", size=3, data=subset(smc.early, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=swc.avg), color="#E69F00", size=3, data=subset(smc.early, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=swc.avg), data=smc.early, lwd=1.4, color="black") + ylab("Soil moisture at emergence (%)") + xlab("Aridity Gradient")+ annotate("text", x=-4.5,y=30, label="c", size=8)
hobo.means <- HOBO.data %>% group_by(season, Microsite, Site) %>% summarize(temp.avg=mean(Temp),temp.var=var(Temp),temp.sd=sd(Temp),temp.se=se(Temp),rh.avg=mean(RH, na.rm=T),rh.var=var(RH, na.rm=T),rh.se=se(RH),frost=(sum(na.omit(Temp)<0)/length(na.omit(Temp))*100),heat=(sum(na.omit(Temp)>30)/length(na.omit(Temp))*100))
hobo.means[,"Year"] <- c(rep(2016,14),rep(2017,14))
hobo.arid <- merge(hobo.means, gams.data, by=c("Site","Year"))
# hobo.arid[,"CV"] <-hobo.arid[,"temp.sd"] /hobo.arid[,"CV"]
## temperature variability
m7 <- lmer(temp.var~Microsite * poly(gams,2) + (1|Year), data=hobo.arid)
shapiro.test(resid(m7))##
## Shapiro-Wilk normality test
##
## data: resid(m7)
## W = 0.96514, p-value = 0.4578anova(m7)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 4406.2 4406.2 1 21.000 21.4200 0.0001448 ***
## poly(gams, 2) 4126.7 2063.3 2 21.483 10.0306 0.0008384 ***
## Microsite:poly(gams, 2) 56.2 28.1 2 21.000 0.1366 0.8730834
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m7) ## R squared values ## R2m R2c
## 0.4521376 0.7930449p7 <- ggplot(hobo.arid) + geom_jitter(aes(x=gams, y=temp.var), color="#56B4E9", size=3, data=subset(hobo.arid, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=temp.var), color="#E69F00", size=3, data=subset(hobo.arid, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=temp.var), color="#E69F00", fill="#E69F0080", data=subset(hobo.arid, Microsite=="open"), lwd=1.4) + stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=temp.var), color="#56B4E9", fill="#56B4E970", data=subset(hobo.arid, Microsite=="shrub"), lwd=1.4) + ylab("Temperature variation") + xlab("Aridity Gradient")+ annotate("text", x=-4.5,y=125, label="d", size=8)
shrubz <- read.csv("Data/ERG.shrub.csv")
shrub.mean <- shrubz %>% group_by(Site, Microsite) %>% summarize(compact=mean(Compaction))
shrub.clim <- data.frame(shrub.mean, gams=rep(gams.data[gams.data$Year==2016,"gams"],each=2))
#shrub.clim <- subset(shrub.clim, Year==2016)
m8 <- lm(log(compact) ~ Microsite * gams, data=shrub.clim)
summary(m8)##
## Call:
## lm(formula = log(compact) ~ Microsite * gams, data = shrub.clim)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.37348 -0.12596 0.04006 0.12843 0.33669
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.138777 0.804662 -0.172 0.8665
## Micrositeshrub -3.101869 1.137963 -2.726 0.0213 *
## gams 0.007791 0.010038 0.776 0.4556
## Micrositeshrub:gams 0.036028 0.014196 2.538 0.0295 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2383 on 10 degrees of freedom
## Multiple R-squared: 0.6967, Adjusted R-squared: 0.6058
## F-statistic: 7.658 on 3 and 10 DF, p-value: 0.006anova(m8)## Analysis of Variance Table
##
## Response: log(compact)
## Df Sum Sq Mean Sq F value Pr(>F)
## Microsite 1 0.18836 0.18836 3.3176 0.098548 .
## gams 1 0.75039 0.75039 13.2168 0.004571 **
## Microsite:gams 1 0.36569 0.36569 6.4409 0.029469 *
## Residuals 10 0.56776 0.05678
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m8) ## R squared values ## R2m R2c
## 0.6386405 0.6386405p8 <- ggplot(shrub.clim) + geom_jitter(aes(x=gams, y=compact), color="#56B4E9", size=3, data=subset(shrub.clim, Microsite=="shrub"))+ geom_jitter(aes(x=gams, y=compact), color="#E69F00", size=3, data=subset(shrub.clim, Microsite=="open"))+ stat_smooth(method="lm", formula= y~exp(x),aes(x=gams, y=compact), color="#56B4E9", fill="#56B4E980", data=subset(shrub.clim, Microsite=="shrub"), lwd=1.4) +
stat_smooth(method="lm", formula= y~exp(x),aes(x=gams, y=compact), color="#E69F00", fill="#E69F0080",data=subset(shrub.clim, Microsite=="open"), lwd=1.4)+ ylab("Soil compaction") + xlab("")+ theme_Publication() + annotate("text", x=-4.4,y=2.5, label="b", size=8) +coord_cartesian(ylim=c(0.5, 2.5))
grid.arrange(p5,p8,p6,p7)
Microsite comparisons
nut <- read.csv("Data//ERG.soilnutrients.csv")
fit1 <- aovp(N ~ microsite, data=nut)## [1] "Settings: unique SS "summary(fit1)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## microsite 1 1031.1 1031.1 5000 0.0014 **
## Residuals 68 6582.7 96.8
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1t.test(N ~ microsite, data=nut)##
## Welch Two Sample t-test
##
## data: N by microsite
## t = -3.2637, df = 36.403, p-value = 0.002396
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -12.444151 -2.907849
## sample estimates:
## mean in group open mean in group shrub
## 1.594571 9.270571fit2 <- aovp(P ~ microsite, data=nut)## [1] "Settings: unique SS "t.test(P ~ microsite, data=nut)##
## Welch Two Sample t-test
##
## data: P by microsite
## t = -2.2393, df = 53.804, p-value = 0.02929
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -7.094126 -0.391588
## sample estimates:
## mean in group open mean in group shrub
## 6.365714 10.108571summary(fit2)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## microsite 1 245.2 245.157 5000 0.0106 *
## Residuals 68 3324.4 48.889
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1fit3 <- aovp(K ~ microsite, data=nut)## [1] "Settings: unique SS "t.test(K ~ microsite, data=nut)##
## Welch Two Sample t-test
##
## data: K by microsite
## t = -2.8179, df = 55.964, p-value = 0.006669
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -154.32488 -26.07512
## sample estimates:
## mean in group open mean in group shrub
## 141.4571 231.6571summary(fit3)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## microsite 1 142381 142381 5000 0.003 **
## Residuals 68 1219331 17931
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1compact <- read.csv("Data//ERG.shrub.csv")
fit4 <- aovp(Compaction ~ Microsite, data=compact)## [1] "Settings: unique SS "t.test(Compaction ~ Microsite, data=compact)##
## Welch Two Sample t-test
##
## data: Compaction by Microsite
## t = 2.6675, df = 411.15, p-value = 0.007943
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## 0.06357828 0.41975506
## sample estimates:
## mean in group open mean in group shrub
## 1.653571 1.411905summary(fit4)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## Microsite 1 6.13 6.1323 5000 0.0044 **
## Residuals 418 360.23 0.8618
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1soilmoist <- read.csv("Data//ERG.phytometer.census.csv")
fit5 <- aovp(swc ~ Microsite, data=subset(soilmoist, Census=="emergence"))## [1] "Settings: unique SS "summary(fit5)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## Microsite 1 46 45.967 179 0.3631
## Residuals 838 62917 75.080t.test(swc ~ Microsite, data=subset(soilmoist, Census=="emergence"))##
## Welch Two Sample t-test
##
## data: swc by Microsite
## t = 0.78246, df = 827.4, p-value = 0.4342
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -0.7057865 1.6415008
## sample estimates:
## mean in group open mean in group shrub
## 11.16500 10.69714micro <- read.csv("Data//ERG.logger.data.csv")
micro.avg <- micro %>% group_by(Year, Site, Microsite, Rep) %>% summarize(temp=mean(Temp), rh=mean(RH))
fit6 <- aovp(temp ~ Microsite, data=micro.avg)## [1] "Settings: unique SS "summary(fit6)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## Microsite 1 77.61 77.608 5000 < 2.2e-16 ***
## Residuals 111 903.55 8.140
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1t.test(temp ~ Microsite, data=micro.avg)##
## Welch Two Sample t-test
##
## data: temp by Microsite
## t = 3.0867, df = 110.64, p-value = 0.002558
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## 0.5933932 2.7216712
## sample estimates:
## mean in group open mean in group shrub
## 11.101469 9.443937fit7 <- aovp(rh ~ Microsite, data=micro.avg)## [1] "Settings: unique SS "summary(fit7)## Component 1 :
## Df R Sum Sq R Mean Sq Iter Pr(Prob)
## Microsite 1 375 375.35 114 0.4737
## Residuals 108 34999 324.07t.test(rh ~ Microsite, data=micro.avg)##
## Welch Two Sample t-test
##
## data: rh by Microsite
## t = -1.0736, df = 105.97, p-value = 0.2854
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
## -10.523642 3.129837
## sample estimates:
## mean in group open mean in group shrub
## 55.04511 58.74201Plot level GLMM
m1 <- glmer.nb(Abundance ~ Microsite * gams + (1|Year), data=community.arid)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?;Model is nearly unidentifiable: large eigenvalue ratio
## - Rescale variables?## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?;Model is nearly unidentifiable: large eigenvalue ratio
## - Rescale variables?## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: very large eigenvalue
## - Rescale variables?car::Anova(m1, type=2)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: Abundance
## Chisq Df Pr(>Chisq)
## Microsite 16.3613 1 5.234e-05 ***
## gams 1136.3911 1 < 2.2e-16 ***
## Microsite:gams 1.0278 1 0.3107
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1coef(m1)## $Year
## (Intercept) Micrositeshrub gams Micrositeshrub:gams
## 2016 13.13187 0.6544052 -0.1298036 -0.005782823
## 2017 11.53434 0.6544052 -0.1298036 -0.005782823
##
## attr(,"class")
## [1] "coef.mer"ggplot(community.arid) + geom_jitter(aes(x=gams, y=Abundance), width = 0.5, color="#56B4E9", size=3, data=subset(community.arid, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=Abundance),width = 0.5, color="#E69F00", size=3, data=subset(community.arid, Microsite=="open")) + theme_Publication() + stat_function(fun = function(x) exp(12.332+ x*-0.1298), lwd=2, lty=1, col="Grey40")+ ylab("annual plant abundance") + xlab("Gams index of continentality")## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
source("rsquared.r")
m2 <- glmer.nb(Richness ~ Microsite * poly(gams,2) + (1|Year), data=community.arid)## Warning in theta.ml(Y, mu, weights = object@resp$weights, limit = limit, :
## iteration limit reachedcar::Anova(m2, type=3)## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: Richness
## Chisq Df Pr(>Chisq)
## (Intercept) 140.6109 1 < 2.2e-16 ***
## Microsite 8.7643 1 0.003072 **
## poly(gams, 2) 969.9903 2 < 2.2e-16 ***
## Microsite:poly(gams, 2) 374.2555 2 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# rsquared.glmm(m2)
newdf <- expand.grid(gams=seq(45,90, by=0.05), Microsite=c("shrub","open"))
newdf$fit = predict(m2, type = "response", newdata = newdf)## Warning in pred + base::drop(as(newRE$b %*% newRE$Zt, "matrix")): longer
## object length is not a multiple of shorter object length## richness
p1 <- ggplot(community.arid) + geom_jitter(aes(x=gams, y=Richness), color="#181818", size=3, alpha=0.4, width = 0.75,shape = 1, data=subset(community.arid, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=Richness), color="#181818", size=2,alpha=0.4, data=subset(community.arid, Microsite=="open"),shape = 2, width = 0.75) + theme_Publication() + ylab("species richness") + xlab("Gams index of continentality")+
geom_smooth(method="glm", method.args=list(family="poisson"), formula= y~poly(x,2),aes(x=gams, y=Richness), color="#181818", fill="#80808080" , data=subset(community.arid, Microsite=="shrub"), lwd=2) +
geom_smooth(method="glm", method.args=list(family="poisson"), formula= y~poly(x,2),aes(x=gams, y=Richness), color="#181818", fill="#80808080", data=subset(community.arid, Microsite=="open"), lwd=2, lty=2)+ ylab("Species richness") + xlab("") + annotate("text", x=50,y=6, label="a", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing#+ geom_smooth(data=subset(newdf,Microsite=="shrub"), aes(x=gams, y=fit), color="#181818",lwd=2)+ geom_smooth(data=subset(newdf,Microsite=="open"), aes(x=gams, y=fit),color="#505050",lwd=2)
m3<- lmer(ihs(Biomass) ~ Microsite * poly(gams,2) + (1|Year), data=community.arid)
anova(m3)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 64.53 64.527 1 833.00 210.49 < 2.2e-16 ***
## poly(gams, 2) 511.61 255.805 2 833.44 834.46 < 2.2e-16 ***
## Microsite:poly(gams, 2) 32.08 16.041 2 833.00 52.33 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m3) ## R squared values## R2m R2c
## 0.5031296 0.8793048## biomass
p2 <- ggplot(community.arid) + geom_jitter(aes(x=gams, y=ihs(Biomass)), color="#181818", size=3, width = 0.75,shape = 1, alpha=0.5, data=subset(community.arid, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=ihs(Biomass)), color="#181818", size=2,shape = 2,alpha=0.5, width = 0.75, data=subset(community.arid, Microsite=="open")) + theme_Publication() +
geom_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=ihs(Biomass)), color="#181818", fill="#80808080", data=subset(community.arid, Microsite=="shrub"), lwd=2) +
geom_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=ihs(Biomass)), color="#181818", fill="#80808080", data=subset(community.arid, Microsite=="open"), lwd=2, lty=2)+ ylab("Annual biomass") + xlab("")+ annotate("text", x=50,y=5, label="b", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacingm4<- lmer(mpd~ Microsite * gams + (1|Year), data=mpd.gams)
anova(m4)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 0.68094 0.68094 1 22 9.2604 0.005967 **
## gams 0.27407 0.27407 1 22 3.7272 0.066522 .
## Microsite:gams 0.55289 0.55289 1 22 7.5189 0.011898 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m4) ## R squared values## R2m R2c
## 0.3852134 0.3852134newdf <- expand.grid(gams=seq(75,90, by=0.05), Microsite=c("shrub","open"))
newdf$fit = predict(m4, type = "response", newdata = newdf)
## phylogenetic diversity
p3 <- ggplot(mpd.gams)+ geom_jitter(aes(x=gams, y=mpd), color="#181818", size=3,shape=1, alpha=0.4, data=subset(mpd.gams, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=mpd), color="#181818",alpha=0.4, size=2,shape=2, data=subset(mpd.gams, Microsite=="open")) + stat_smooth(method="lm", formula= y~x,aes(x=gams, y=mpd), data=subset(mpd.gams, Microsite=="shrub"), color="#181818",fill="#80808080", lwd=2)+ stat_smooth(method="lm", formula= y~x,aes(x=gams, y=mpd), data=subset(mpd.gams, Microsite=="open"), color="#181818",fill="#80808080", lwd=2, lty=2)+ ylab("Phylogenetic community dissimilarity") + xlab("Gams index of continentality")+ theme_Publication()+ annotate("text", x=50,y=1.4, label="c", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## native vs non-native
status.gams <- merge(status.data, gams.data, by=c("Year","Site"))
m5<- glmer.nb(native~ Microsite * scale(gams) + (1|Year), data=status.gams) ## need to scale data## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control
## $checkConv, : Model failed to converge with max|grad| = 0.00214781 (tol =
## 0.001, component 1)car::Anova(m5, type=3)## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: native
## Chisq Df Pr(>Chisq)
## (Intercept) 565.724 1 < 2.2e-16 ***
## Microsite 23.122 1 1.521e-06 ***
## scale(gams) 67.393 1 2.224e-16 ***
## Microsite:scale(gams) 35.708 1 2.292e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# rsquared.glmm(m5)
m6<- glmer.nb(non.native~ Microsite * scale(gams) + (1|Year), data=status.gams) ## need to scale data
car::Anova(m6, type=3)## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: non.native
## Chisq Df Pr(>Chisq)
## (Intercept) 12.5323 1 0.000400 ***
## Microsite 13.5445 1 0.000233 ***
## scale(gams) 353.8669 1 < 2.2e-16 ***
## Microsite:scale(gams) 1.4368 1 0.230664
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# rsquared.glmm(m5)
status.plot <- gather(status.gams, origin, abundance, 5:6)
p4 <- ggplot(subset(status.plot, abundance>0)) + geom_jitter(aes(x=gams, y=abundance), color="#181818", size=3, data=subset(status.plot, Microsite=="shrub"), width = 1, shape=1, alpha=0.3) +geom_jitter(aes(x=gams, y=abundance), width = 1, color="#181818", size=2,alpha=0.3,shape=2, data=subset(status.plot, Microsite=="open")) + theme_Publication() +
geom_smooth(method="glm", method.args=list(family="poisson"), formula= y~x,aes(x=gams, y=abundance), color="#505050", se=F, data=subset(status.plot, Microsite=="open" & origin=="native"), lwd=2, lty=2)+
geom_smooth(method="glm", method.args=list(family="poisson"), formula= y~x,aes(x=gams, y=abundance),se=F, color="#181818", data=subset(status.plot, Microsite=="open" & origin=="non.native"), lwd=2, lty=2)+
geom_smooth(method="glm", method.args=list(family="poisson"), formula= y~x,aes(x=gams, y=abundance),se=F, color="#505050", data=subset(status.plot, Microsite=="shrub" & origin=="native"), lwd=2)+
geom_smooth(method="glm", method.args=list(family="poisson"), formula= y~x,aes(x=gams, y=abundance),se=F, color="#181818", data=subset(status.plot, Microsite=="shrub" & origin=="non.native"), lwd=2) + ylab("Annual plant abundance") + xlab("Gams index of continentality")+ annotate("text", x=50,y=310, label="d", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacinggrid.arrange(p1,p2,p3,p4)## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
### Nutrients
nutrient.mean <- nutrients.climate %>% group_by(Site, arid.gradient, microsite) %>% summarize(nitrogen=mean(N), phosphorus=mean(P), potassium=mean(K))
nutrient.mean <- data.frame(nutrient.mean,gams=rep(gams.data[gams.data$Year==2016,"gams"],each=2))
## nitrogen
m5 <- lm(nitrogen ~ poly(gams,2)* microsite, data=nutrient.mean)
summary(m5)##
## Call:
## lm(formula = nitrogen ~ poly(gams, 2) * microsite, data = nutrient.mean)
##
## Residuals:
## Min 1Q Median 3Q Max
## -3.2918 -0.9251 -0.2773 0.8012 5.6374
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.595 1.038 1.536 0.163115
## poly(gams, 2)1 3.927 3.885 1.011 0.341676
## poly(gams, 2)2 3.847 3.885 0.990 0.350969
## micrositeshrub 7.676 1.468 5.228 0.000795 ***
## poly(gams, 2)1:micrositeshrub 22.258 5.494 4.052 0.003676 **
## poly(gams, 2)2:micrositeshrub 17.833 5.494 3.246 0.011771 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2.747 on 8 degrees of freedom
## Multiple R-squared: 0.9298, Adjusted R-squared: 0.8859
## F-statistic: 21.18 on 5 and 8 DF, p-value: 0.0002012anova(m5)## Analysis of Variance Table
##
## Response: nitrogen
## Df Sum Sq Mean Sq F value Pr(>F)
## poly(gams, 2) 2 389.59 194.794 25.817 0.0003239 ***
## microsite 1 206.22 206.223 27.332 0.0007948 ***
## poly(gams, 2):microsite 2 203.35 101.677 13.476 0.0027446 **
## Residuals 8 60.36 7.545
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m5) ## R squared values ## R2m R2c
## 0.8906817 0.8906817p5 <- ggplot(nutrient.mean) + geom_jitter(aes(x=gams, y=nitrogen), color="#181818", size=3, shape=16,alpha=0.5, data=subset(nutrient.mean, microsite=="shrub"))+ geom_jitter(aes(x=gams, y=nitrogen), color="#181818", size=3,shape=17, alpha=0.4, data=subset(nutrient.mean, microsite=="open"))+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=nitrogen), color="#181818", fill="#80808080", data=subset(nutrient.mean, microsite=="shrub"), lwd=2) +
stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=nitrogen), color="#181818", fill="#80808080",data=subset(nutrient.mean, microsite=="open"), lwd=2, lty=2)+ ylab("Nitrogen (ppm)") + xlab("")+ theme_Publication()+ coord_cartesian(ylim=c(-2, 26)) + xlab("")+ annotate("text", x=63,y=25, label="a", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacingsmc.early <- subset(smc, Census=="emergence")
smc.early <- merge(smc.early, gams.data, by=c("Year","Site"))
## nitrogen
m6 <- lmer(log(swc.avg) ~ poly(gams,2)* Microsite + (1|Year), data=smc.early)
summary(m6)## Linear mixed model fit by REML t-tests use Satterthwaite approximations
## to degrees of freedom [lmerMod]
## Formula: log(swc.avg) ~ poly(gams, 2) * Microsite + (1 | Year)
## Data: smc.early
##
## REML criterion at convergence: 16.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.75790 -0.67057 0.05989 0.64573 1.52776
##
## Random effects:
## Groups Name Variance Std.Dev.
## Year (Intercept) 0.32730 0.5721
## Residual 0.09642 0.3105
## Number of obs: 28, groups: Year, 2
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 2.183930 0.412962 1.000000 5.288
## poly(gams, 2)1 -4.095881 0.529751 21.483000 -7.732
## poly(gams, 2)2 0.009355 0.444206 21.040000 0.021
## Micrositeshrub -0.021937 0.117363 21.000000 -0.187
## poly(gams, 2)1:Micrositeshrub 0.108296 0.621024 21.000000 0.174
## poly(gams, 2)2:Micrositeshrub -0.128193 0.621024 21.000000 -0.206
## Pr(>|t|)
## (Intercept) 0.119
## poly(gams, 2)1 1.22e-07 ***
## poly(gams, 2)2 0.983
## Micrositeshrub 0.854
## poly(gams, 2)1:Micrositeshrub 0.863
## poly(gams, 2)2:Micrositeshrub 0.838
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) pl(,2)1 pl(,2)2 Mcrsts p(,2)1:
## ply(gms,2)1 0.000
## ply(gms,2)2 0.000 0.084
## Microstshrb -0.142 0.000 0.000
## ply(g,2)1:M 0.000 -0.586 0.000 0.000
## ply(g,2)2:M 0.000 0.000 -0.699 0.000 0.000anova(m6)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## poly(gams, 2) 8.6913 4.3456 2 21.339 45.071 2.183e-08 ***
## Microsite 0.0034 0.0034 1 21.000 0.035 0.8535
## poly(gams, 2):Microsite 0.0070 0.0035 2 21.000 0.037 0.9642
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m6) ## R squared values ## R2m R2c
## 0.5883174 0.9063213p6 <- ggplot(smc.early) + geom_jitter(aes(x=gams, y=swc.avg), color="#181818", size=3, shape=16,alpha=0.5, data=subset(smc.early, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=swc.avg), color="#181818", size=3,shape=17,alpha=0.5, data=subset(smc.early, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=swc.avg), data=smc.early, lwd=2, color="black") + ylab("Soil moisture at emergence (%)") + xlab("")+ annotate("text", x=50,y=35, label="b", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## temperature variability
m7 <- lmer(temp.var~Microsite * poly(gams,2) + (1|Year), data=hobo.arid)
shapiro.test(resid(m7))##
## Shapiro-Wilk normality test
##
## data: resid(m7)
## W = 0.96514, p-value = 0.4578anova(m7)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value Pr(>F)
## Microsite 4406.2 4406.2 1 21.000 21.4200 0.0001448 ***
## poly(gams, 2) 4126.7 2063.3 2 21.483 10.0306 0.0008384 ***
## Microsite:poly(gams, 2) 56.2 28.1 2 21.000 0.1366 0.8730834
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m7) ## R squared values ## R2m R2c
## 0.4521376 0.7930449p7 <- ggplot(hobo.arid) + geom_jitter(aes(x=gams, y=temp.var), color="#181818", size=3, shape=16, alpha=0.5, data=subset(hobo.arid, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=temp.var), color="#181818", size=3, shape=17, alpha=0.5, data=subset(hobo.arid, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=temp.var), color="#181818", fill="#80808080", data=subset(hobo.arid, Microsite=="open"), lwd=2) + stat_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=temp.var), color="#181818", fill="#80808080", data=subset(hobo.arid, Microsite=="shrub"), lwd=2, lty=2) + ylab("Temperature variation") + xlab("Gams index of continentality")+ annotate("text", x=50,y=125, label="c", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## soil compaction
m8 <- lm(log(compact) ~ Microsite * gams, data=shrub.clim)
summary(m8)##
## Call:
## lm(formula = log(compact) ~ Microsite * gams, data = shrub.clim)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.37348 -0.12596 0.04006 0.12843 0.33669
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.138777 0.804662 -0.172 0.8665
## Micrositeshrub -3.101869 1.137963 -2.726 0.0213 *
## gams 0.007791 0.010038 0.776 0.4556
## Micrositeshrub:gams 0.036028 0.014196 2.538 0.0295 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.2383 on 10 degrees of freedom
## Multiple R-squared: 0.6967, Adjusted R-squared: 0.6058
## F-statistic: 7.658 on 3 and 10 DF, p-value: 0.006anova(m8)## Analysis of Variance Table
##
## Response: log(compact)
## Df Sum Sq Mean Sq F value Pr(>F)
## Microsite 1 0.18836 0.18836 3.3176 0.098548 .
## gams 1 0.75039 0.75039 13.2168 0.004571 **
## Microsite:gams 1 0.36569 0.36569 6.4409 0.029469 *
## Residuals 10 0.56776 0.05678
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m8) ## R squared values ## R2m R2c
## 0.6386405 0.6386405p8 <- ggplot(shrub.clim) + geom_jitter(aes(x=gams, y=log(compact)), color="#181818", size=3, shape=16, alpha=0.5, data=subset(shrub.clim, Microsite=="shrub")) +geom_jitter(aes(x=gams, y=log(compact)), color="#181818", size=3, shape=17, alpha=0.5, data=subset(shrub.clim, Microsite=="open")) + theme_Publication()+ stat_smooth(method="lm", formula= y~x, aes(x=gams, y=log(compact)), color="#181818", fill="#80808080", data=subset(shrub.clim, Microsite=="open"), lwd=2) + stat_smooth(method="lm", formula= y~x,aes(x=gams, y=log(compact)), color="#181818", fill="#80808080", data=subset(shrub.clim, Microsite=="shrub"), lwd=2, lty=2) + ylab("Soil compaction") + xlab("Gams index of continentality")+ annotate("text", x=63,y=1.1, label="d", size=8) ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacinggrid.arrange(p5,p6,p7,p8)## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
Phytometer with GAMS
## treatment plots
census.long <- gather(census.arid, species, biomass, 12:14)
census.long[,"species"] <- ifelse(census.long[,"species"] =="Phacelia.biomass", "P. tanacetifolia",
ifelse(census.long[,"species"]=="Plantago.biomass","P. insularis","S. columbariae"))
census.long[,"species"] <- as.factor(census.long$species)
## microsite average
census.micro <- census.long %>% filter(biomass>0) %>% group_by(Microsite, species) %>% summarize(Biomass=mean(biomass), error=se(biomass))
p9 <- ggplot(census.micro, aes(x=species, y=Biomass, fill=Microsite)) + geom_bar(position=position_dodge(), stat="identity", colour="black") + scale_fill_brewer(palette="Greys")+ geom_errorbar(aes(ymin=Biomass-error, ymax=Biomass+error),
width=.2, # Width of the error bars
position=position_dodge(.9))+ theme_Publication() + theme(axis.text.x = element_text(size=16, face="italic"),axis.text.y = element_text(size=16),axis.title = element_text(size=20),legend.text=element_text(size=18))## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## nutrient average
census.nut <- census.long %>% filter(biomass>0) %>% group_by(Nutrient, species) %>% summarize(Biomass=mean(biomass), error=se(biomass))
p10 <- ggplot(census.nut, aes(x=species, y=Biomass, fill=Nutrient)) + geom_bar(position=position_dodge(), stat="identity",colour="black") + scale_fill_brewer(palette="Greys")+ geom_errorbar(aes(ymin=Biomass-error, ymax=Biomass+error),
width=.2, # Width of the error bars
position=position_dodge(.9))+ theme_Publication()+ theme(axis.text.x = element_text(size=16, face="italic"),axis.text.y = element_text(size=16),axis.title = element_text(size=20),legend.text=element_text(size=18))## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacing## aridity average
# ggplot(census.long, aes(x=arid.gradient, y=biomass, fill=species, colour=species)) + theme_Publication()+ stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=biomass), color="#009E73", fill="#009E7340", data=subset(census.long, species=="Phacelia.biomass" & biomass>0)) + stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=biomass), color="#505050",fill="#50505040", data=subset(census.long, species=="Plantago.biomass" & biomass>0))+ stat_smooth(method="lm", formula= y~exp(x),aes(x=arid.gradient, y=biomass), color="#181818", fill="#18181840", data=subset(census.long, species=="Salvia.biomass" & biomass>0))
### need to separate zeros using gamma-hurdle model
##http://seananderson.ca/2014/05/18/gamma-hurdle.html
## logistic first
census.long[,"presence"] <- ifelse(census.long$biomass>0,1,0)
census.long <- merge(census.long, gams.data, by=c("Year","Site"))
## Drop panoche 2017 because of cold
census.npan <- census.long %>% filter(Site!="PanocheHills" | Year!="2017")
m1.l <- glmer(presence~Microsite * gams * Nutrient + (1|Year),data=subset(census.npan, species=="P. tanacetifolia"), family=binomial)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, : Model is nearly unidentifiable: large eigenvalue ratio
## - Rescale variables?car::Anova(m1.l, type=2)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: presence
## Chisq Df Pr(>Chisq)
## Microsite 29.2566 1 6.34e-08 ***
## gams 8.9859 1 0.002721 **
## Nutrient 0.0060 1 0.938002
## Microsite:gams 4.4524 1 0.034853 *
## Microsite:Nutrient 1.9750 1 0.159917
## gams:Nutrient 0.0019 1 0.965645
## Microsite:gams:Nutrient 0.0139 1 0.906248
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m1 <- lmer(log(biomass)~Microsite * poly(gams,2) * Nutrient + (1|Year),data=subset(census.long, species=="P. tanacetifolia" & presence==1))
shapiro.test(resid(m1))##
## Shapiro-Wilk normality test
##
## data: resid(m1)
## W = 0.99225, p-value = 0.2595anova(m1)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value
## Microsite 22.121 22.121 1 221 12.213
## poly(gams, 2) 157.099 78.549 2 221 43.368
## Nutrient 5.710 5.710 1 221 3.152
## Microsite:poly(gams, 2) 15.688 7.844 2 221 4.331
## Microsite:Nutrient 0.078 0.078 1 221 0.043
## poly(gams, 2):Nutrient 2.854 1.427 2 221 0.788
## Microsite:poly(gams, 2):Nutrient 1.406 0.703 2 221 0.388
## Pr(>F)
## Microsite 0.0005733 ***
## poly(gams, 2) < 2.2e-16 ***
## Nutrient 0.0771914 .
## Microsite:poly(gams, 2) 0.0142925 *
## Microsite:Nutrient 0.8361223
## poly(gams, 2):Nutrient 0.4561215
## Microsite:poly(gams, 2):Nutrient 0.6786980
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m1) ## R squared values ## R2m R2c
## 0.346015 0.346015m2.l <- glmer(presence~Microsite * poly(gams,2) * Nutrient + (1|Year),data=subset(census.npan, species=="P. insularis"), family=binomial)
car::Anova(m2.l, type=2)## Analysis of Deviance Table (Type II Wald chisquare tests)
##
## Response: presence
## Chisq Df Pr(>Chisq)
## Microsite 1.6804 1 0.194870
## poly(gams, 2) 9.6027 2 0.008219 **
## Nutrient 0.3064 1 0.579874
## Microsite:poly(gams, 2) 5.3976 2 0.067285 .
## Microsite:Nutrient 1.9903 1 0.158314
## poly(gams, 2):Nutrient 3.4679 2 0.176586
## Microsite:poly(gams, 2):Nutrient 0.6252 2 0.731534
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m2 <- lmer(log(biomass)~Microsite * poly(gams,2) * Nutrient + (1|Year), data=subset(census.long, species=="P. insularis" & presence==1))
shapiro.test(resid(m2))##
## Shapiro-Wilk normality test
##
## data: resid(m2)
## W = 0.99193, p-value = 0.5364anova(m2)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value
## Microsite 0.569 0.5693 1 141.23 0.5147
## poly(gams, 2) 36.153 18.0765 2 137.31 16.3429
## Nutrient 8.958 8.9576 1 141.27 8.0986
## Microsite:poly(gams, 2) 0.858 0.4289 2 141.51 0.3878
## Microsite:Nutrient 0.110 0.1096 1 141.02 0.0991
## poly(gams, 2):Nutrient 2.373 1.1866 2 141.20 1.0728
## Microsite:poly(gams, 2):Nutrient 0.654 0.3272 2 141.03 0.2958
## Pr(>F)
## Microsite 0.47430
## poly(gams, 2) 4.297e-07 ***
## Nutrient 0.00509 **
## Microsite:poly(gams, 2) 0.67929
## Microsite:Nutrient 0.75338
## poly(gams, 2):Nutrient 0.34481
## Microsite:poly(gams, 2):Nutrient 0.74439
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m2) ## R squared values ## R2m R2c
## 0.2802814 0.5695942m3.l <- glmer(presence~Microsite * poly(gams,2) * Nutrient + (1|Year),data=subset(census.npan, species=="S. columbariae"), family=binomial)
car::Anova(m3.l, type=3)## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: presence
## Chisq Df Pr(>Chisq)
## (Intercept) 2.1914 1 0.13878
## Microsite 1.6046 1 0.20525
## poly(gams, 2) 25.8433 2 2.445e-06 ***
## Nutrient 0.1181 1 0.73109
## Microsite:poly(gams, 2) 1.8388 2 0.39875
## Microsite:Nutrient 3.3815 1 0.06593 .
## poly(gams, 2):Nutrient 0.9600 2 0.61878
## Microsite:poly(gams, 2):Nutrient 0.4124 2 0.81369
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1m3 <- lmer(log(biomass)~Microsite * poly(gams,2) * Nutrient + (1|Year), data=subset(census.long, species=="S. columbariae" & biomass>0))
shapiro.test(resid(m3))##
## Shapiro-Wilk normality test
##
## data: resid(m3)
## W = 0.99363, p-value = 0.1532anova(m3)## Analysis of Variance Table of type III with Satterthwaite
## approximation for degrees of freedom
## Sum Sq Mean Sq NumDF DenDF F.value
## Microsite 0.350 0.3503 1 333.03 0.2070
## poly(gams, 2) 61.028 30.5139 2 327.98 18.0272
## Nutrient 18.797 18.7973 1 333.12 11.1052
## Microsite:poly(gams, 2) 1.304 0.6521 2 333.03 0.3853
## Microsite:Nutrient 0.386 0.3859 1 333.05 0.2280
## poly(gams, 2):Nutrient 2.914 1.4569 2 333.05 0.8607
## Microsite:poly(gams, 2):Nutrient 3.356 1.6780 2 333.01 0.9913
## Pr(>F)
## Microsite 0.6494522
## poly(gams, 2) 3.733e-08 ***
## Nutrient 0.0009576 ***
## Microsite:poly(gams, 2) 0.6805772
## Microsite:Nutrient 0.6333374
## poly(gams, 2):Nutrient 0.4237872
## Microsite:poly(gams, 2):Nutrient 0.3721737
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r.squaredGLMM(m3) ## R squared values ## R2m R2c
## 0.1506253 0.2337409## summary presence
mean.occ <- census.npan %>% group_by(gams,species, Microsite) %>% summarize(presence=mean(presence))
p11 <- ggplot(census.npan, aes(x=gams, y=presence, fill=species, color=species)) + theme_Publication()+
geom_smooth(method="glm", method.args=list(family="binomial"), formula= y~x,aes(x=gams, y=presence),fill="#0daec2", color="#0daec2", data=subset(census.npan, species=="P. tanacetifolia" & Microsite=="shrub"), lwd=2)+
geom_smooth(method="glm", method.args=list(family="binomial"), formula= y~x,aes(x=gams, y=presence),fill="#0daec2", color="#0daec2", data=subset(census.npan, species=="P. tanacetifolia"& Microsite=="open"), lwd=2, lty=2) + geom_smooth(method="glm", method.args=list(family="binomial"), formula= y~poly(x,2),aes(x=gams, y=presence), color="#fb6f70", fill="#fb6f7050", data=subset(census.npan, species=="P. insularis"), lwd=2)+ geom_smooth(method="glm", method.args=list(family="binomial"), formula= y~poly(x,2),aes(x=gams, y=presence),fill="#b7d73850", color="#b7d738", data=subset(census.npan, species=="S. columbariae"), lwd=2)+ ylab("Probability of occurrence") + xlab("Gams index of continentality")## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacingp12 <- ggplot(census.long, aes(x=gams, y=biomass, fill=species, color=species)) + theme_Publication() + geom_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=log(biomass)), color="#fb6f70", fill="#fb6f7050", data=subset(census.long, species=="P. insularis" & biomass>0), lwd=2, fullrange=TRUE) + geom_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=log(biomass)),fill="#b7d73850", color="#b7d738", data=subset(census.long, species=="S. columbariae"& biomass>0), lwd=3)+
geom_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=log(biomass)),fill="#0daec2", color="#0daec2", data=subset(census.long, species=="P. tanacetifolia" & biomass>0 & Microsite=="shrub"), lwd=2, fullrange=TRUE)+geom_smooth(method="lm", formula= y~poly(x,2),aes(x=gams, y=log(biomass)),fill="#0daec2", color="#0daec2", data=subset(census.long, species=="P. tanacetifolia" & biomass>0 & Microsite=="open"), lwd=2,lty=2, fullrange=TRUE)+ ylab("Phytometer biomass") + xlab("Gams index of continentality") ## Warning: `legend.margin` must be specified using `margin()`. For the old
## behavior use legend.spacinggrid.arrange(p9,p11,p10, p12)## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database
## Warning in grid.Call(L_textBounds, as.graphicsAnnot(x$label), x$x, x$y, :
## font family not found in Windows font database## Warning in grid.Call.graphics(L_text, as.graphicsAnnot(x$label), x$x, x
## $y, : font family not found in Windows font database
## test to see if GAMS related to aridity
arid.gam <- merge(arid.vals, gams.data, by=c("Site"))
cor(arid.gam$Gradient, arid.gam$gams)## [1] -0.5506378